The Peopling Of Australia

Much attention has been devoted to the question of when and how Aborigines first entered the Australian continent, but very little has been paid to what happened after that. By the time Europeans reached Australia, there were probably more than a million Aborigines in the continent, but how long did it take to build up to this number and when were the various environmental zones first occupied? Three main attempts to answer these questions have been made. The first was by American anthropologist Joseph Birdsell, the second by Sandra Bowdler, and the third by David Horton.


Birdsell's theory involved numerous assumptions about generation length, rate of reproduction and group size. The normal size of a colonising group was taken as twenty-five persons, and it was proposed that new colonising groups hived off when a population had reached 60 per cent of an area's carrying capacity. It was assumed that when Europeans first entered Australia, the Aboriginal population was about 300,000 and that the continent was at maximum carrying capacity. However, the figure of 300,000 for the Aboriginal population of Australia at the time of European settlement is based on very slender evidence and is far too low. Nor can it be taken for granted that the continent was fully occupied during Pleistocene times; this may be a more recent phenomenon. Likewise, population density cannot be assumed to have remained unchanged from the Pleistocene, through a period of major climatic change, to modern times. it is, therefore, rather rash to use people-land ratios of modern Aboriginal groups in a model of ice age migrations.

Birdsell's model, nevertheless, may give at least a rough idea of how long it might have taken to achieve 'maximum saturation' of the Australian continent. His computer simulation gave the specific answer of '2204 years of total elapsed time'. Not only did Birdsell argue for a fairly rapid 'filling-up' of Australia, but he also saw the inland regions as being populated more quickly than the coasts. The reasoning behind this argument is that the drier inland regions could support fewer people than the well-watered coastal zones, so the land's carrying capacity would be reached sooner, and people would move on in search of a new foraging area more often. This theory is testable from both archaeological and environmental evidence, and has been challenged on both grounds by Bowdler. She analysed the archaeological evidence and concluded that it did not support Birdsell's theory. She proposed an alternative hypothesis in 1977, based on archaeological evidence, of a coastal colonisation of Australia. Bowdler's thesis is basically that Australia was colonised by people with an economy adapted to living on tropical coasts, and that the similarities between the edible tropical plants and coastal environments of the new continent and their departed homelands would have facilitated settlement. It is argued that their marine-based technology and economy, with a heavy reliance on scale fish, shellfish and small mammals, required little change to exploit Australian coastal resources, and thus change was minimal for many thousands of years. The new settlers gradually moved around the periphery of the continent and expanded inland along major river and lake systems. It was not until much later, she suggests, that the dry heart of the continent was occupied, when people had developed techniques to hunt unfamiliar terrestrial animals such as large kangaroos and began to use grindstones to exploit grass seeds as food. 

Evidence in support of this coastal, gradual and conservative colonisation, in contrast to Birdswell's raid, radiational model, was the peripheral distribution of Pleistocene archaeological sites in Australia (see figure 2.3). Pleistocene sites tend to be either on or near the coast or lie up major river valleys and the associated lake systems such as the Willandra Lakes in the Murray-Darling basin. A few sites were several hundred kilometres inland, but Bowdler argued that they may still have been reached by following large rivers up from the coast. For example, Kenniff Cave lies on the Great Dividing Range in southern Queensland, 500 kilometres inland from the present coastline, but it is near the headwaters of the Darling River system. A trek of some 1800 kilometres would have been required to reach Kenniff by way of the Murray and the Darling, but dispersal was probably over hundreds or thousands of years.

Models for the peopling of Australia. Points of entry may have been Arnhem Land and/or the Kimberley region and/or Cape York.
On the left is Birdsell's radiational model; in the centre, Bowdler's coastal colonisation model; and on the right, Horton's model.

Bowdler's theory was attractive, but in 1977 little archaeological work had been done in the centre of the continent which might support or refute it; excavation had been concentrated around the coast near the modern centres of population. When so much of Australia remained archaeologically unexplored, it was perhaps premature to try to establish the pattern of colonisation. But hypotheses derived from these models can be tested by field work. If colonisation were coastal, the earliest sites should only found near Pleistocene coasts or major river and lake systems; sites in desert and mountain environments should be later. If, on the other hand, colonisation were radiational and continent-wide, some early sites should be found in inland areas, far from coasts, large rivers and associated lakes. Horton's concept was that people with an adaptable, all-purpose economy moved into Australia through well-watered regions on both sides of the Great Dividing Range and penetrated all but the arid core of the continent by 25,000 years ago. Then, when Australia began to dry up, Horton believed that the megafauna became extinct, and people retreated to the coastal regions, not moving back until the present climate was established about 12,000 years ago. Tindale also believed that initial migration penetrated southwards on both sides of the Dividing Range, but present archaeological evidence supports neither this nor Horton's theory.

The discovery by Michael Smith in 1986 of Pleistocene occupation in the Central Desert has ended over a decade of speculation about the timing of human settlement in the core of the continent. Puritjarra rock-shelter is west of Alice Springs, almost in the dead centre of Australia, and shows that the arid interior was settled extremely early. A date of about 22,000 BP for the lowest artefacts was published in 1987, but further work has now revealed that the base of the occupation dates back much further. Dating is still in progress, but Smith says the age is 'in the order of 30,000 years'. Evidence of 'old' inland sites discovered since 1977 prompted Bowdler to re-examine her coastal colonisation theory. In 1992 she wrote:

The process of adaptation to the new continent now seems to have been considerably more rapid than originally supposed. While it still seems probable that original colonizing routes were around the coasts and thence up the major river systems (Bowdler 1977:205; Bowdler 1990c), it is clear that non-coastal environments were successfully exploited on a regular basis quite early on. The Tasmanian evidence indicates this (Cosgrove et al. 1990), as does recent research in Western Australia. In the Kimberley, the site of Widgingarri I was first occupied at least 28,000 years ago, when it would have been some 50 km from the coast, and continued in occupation until the sea had reached to its maximum extent (O'Connor 1990:342).

In Cape York, Arnhem Land and the Kimberley, Pleistocene occupation has been found in rock-shelters near rivers that flow across plains to the coast. The earliest northern sites lie at least 50 kilometres from the present coast and would have been much further from the shore at times of low sea level. Soundings in the sea have revealed the existence of river valleys, through which the present rivers would have flowed across the now submerged continental shelves, along which these early people would have moved. The first landfalls could have been made anywhere across northern Australia from the Kimberleys in the west, to Arnhem Land in the centre, to Cape York in the east. The oldest occupation yet found in the north is in Arnhem Land, but pre-30,000-years-old occupation has also been found in the Kimberleys and Cape York Peninsula (Sandy Creek Shelter near Laura - 32,000 BP - and Nurrabullgin Cave west of Cairns - 37,000 BP).

The environment on the Sahul shelf encountered by early migrants was probably similar to that of the northern third of the Australian continent today. During the last ice age, mean annual temperatures were several degrees lower than today. From about 50,000 to 30,000 years ago, the climate was cooler and moister. Then, after 30,000 BP, the climate became even colder and effective rainfall declined, leading to a period of maximum cold and aridity about 18,000 years ago. The subsequent improvement accelerated around 12,000 BP, and by 10,000 BP the climate was warm and moist. These favourable conditions gradually changed to the climate of present-day Australia by about 3000 years ago. Increased aridity at times of lower sea level would have caused a coastward migration of the vegetation of arid inland Australia. A drop in sea level of only 80 metres would have been sufficient to turn the whole of the Arafura Sea and the Gulf of Carpentaria into land, probably covered with savanna woodland and swamps. The resulting open woodland on the Sahul shelf would have been a suitable human habitat and facilitated movement, but newcomers w2ould have had to adapt to the marked seasonality of climate in the north, with a long dry season in winter.

Was There an Archaeologically 'Invisible' Phase?

The question of whether there was an archaeologically 'invisible' phase is of course unanswerable, if taken literally, but it addresses the suggestion that there was a phase of 'invisible colonisation' of unknown length not yet represented in the archaeological record. Can our earliest dates be taken 'at face value' or do they represent merely the end point of a long archaeologically invisible phase, the scanty remains of which have not survived the vagaries of time? One of the problems in the search for traces of the earliest colonists is that many of the very earliest sites now probably lie underneath the sea, drowned by the post-glacial rise in sea level. Not until the people had moved 50 or more kilometres inland, to what is now the coastline of northern Australia, would any traces of these first migrants remain. It is most unlikely, therefore, that we will find any evidence of early sea-voyaging, because of submergence of landing places and the unlikelihood of organic remains such as wood surviving for more than forty millennia. However, the move inland away from coastal swamps could have been quite rapid, and on some stretches of the Australian shore the continental shelf is not very wide, so early coastal sites would not necessarily have been later submerged.

Erosion at Lake Mungo, New south Wales, on the Walls of china lunette.
(Australian Information Service)

Absence of artefacts in deposits dated to more than 60,000 years also argues against an earlier invisible phase. In the Willandra Lakes, cultural material is totally absent from the lowest sediments, which lie underneath the artefact-bearing Mungo unit, yet thee were earlier periods when the lake contained fresh water and would have been a fafourable human habitat. Similarly, sand dunes on the eastern coast from the same period are devoid of any evidence of human occupation. Moreover, in the Arnhem Land sites described below, sterile sands lie beneath the lowest artefacts.

Koonalda Cave, Nullarbor Plain, South Australia. (D.J. Mulvaney)! Shop Mizuno Team Sports! Never Settle!

Now - $50 Off the WINDLITE™ FZ FLEECE at! Limited Time Only!

539202_Free Ground Shipping 88x31

Budget DVD Sale

The Distribution Of Pleistocene Sites

The speed of new discoveries has been astonishing. There are now more than 170 Pleistocene sites in Australia dated to more than 10,000 years. This enables us to draw finer distinctions between earlier and later Pleistocene sites, without putting too much reliance on precise dates. The earliest sites in Australia, giving dates around or in excess of 40,000 BP, have been found in Arnhem Land (as described below), southwestern Australia (Upper Swan), western New South Wales (Mungo) and southeastern Australia (Cranebrook, near Sydney-but the Cranebrook evidence is rejected as dubious by many). These sites all fit with Bowdler's theory that the first colonisation was essentially coastal.

Top: Bone bead from Devil's Lair, Western Australia, from a 15,000-year-old layer. (C. Dortch)
Centre: Karton tools from South Australia: horsehoof core and pebble tools.
Bottom: Necklace made from Conus species shell beads from a 32,000-year-old layer in Mandu Mandu Creek rock-shelter, Western Australia.
(Hypothetical arrangement of the twenty-two beads by Kate Morse; photo by Douglas Elford, Western Australian Museum)

By about 35,000 years BP or earlier, people were occupying most major ecological zones: the tropical coasts such as the Pilbara (Mandu Mandu Creek rock-shelter, 34,000 BP) and Cape York Peninsula (Ngarabullgan Cave near Cairns, 37,000 BP); the arid zones such as the Nullarbor Plain (Allen's Cave, with a preliminary TL date of c.35,000 BP; the semiarid zone of the Wilandra Lakes; and caves in the highlands of Tasmania such as Warreen Cave (35,000 BP) in the Maxwell River Valley.

By 30,000 years ago, the Pleistocene inhabitant were not only settled on much of the coast of Australia, at Devil's Lair in the west and Keilor in the south, but had also penetrated to the heart of Australia's desert core - to Puritjarra, by a permanent rock pool in the Cleland Hills. By 35,000 BP, there are Pleistocene sites in virtually every ecological zone.

Below let us examine the Arnhem Land sites and dating methods involved in the claim of a 50,000 to 60,000 years antiquity. Other significant Pleistoce sites in the north and south of the continent and Tasmania will be discussed in the following chapters.

The Radiocarbon Bottleneck

A potential solution to this radiocarbon bottleneck was provided by the development of thermoluminescence (TL) dating of naturally deposited sands. This new method dates the time since the artefact-bearing quartz sand was last exposed to sunlight. It has been explained as follows by Bert Roberts and Nigel Spooner, who have been involved in the method's application at key early human occupation sites in Arnhem Land and the Nullarbor Plain:

The light emitted by a mineral upon heating is called thermoluminescence (TL) while optically-stimulated luminescence (OSL) refers to the luminescence emitted by a mineral when exposed to visible (e.g. green laser) light. Luminescence dating comprises the allied techniques of TL and optical dating, both capable of providing calendar years using minerals such as quartz. The upper age limit of luminescence dating for Australian quarts is typically 200,000 years, hence enabling absolute ages to be obtained for archaeological deposits beyond the radiocarbon 'barrier' of 30,000-40,000 years...

In the luminescence dating of sediments, the luminescence 'clock' is reset by sunlight and 'time zero' is the moment when the sediment is buried. However, quartz always retains a small 'residual' TL signal from the moment of burial, and determining this residual level is critical to obtaining reliable TL dates. Optical dating avoids the problem of the residual level by using only the most light-sensitive signal, which can be expected to have been fully reset at the time of burial. (The TL method dates the time since the sand was last exposed to sunlight.) Following burial, the luminescence signal increases steadily with time as a result of the energy absorbed from ionizing radiation (emanating principally from the radio-active decay of naturally-occurring isotopes in soil and rock). The luminescence signal acquired during burial is measured in the laboratory when the sediment is either heated (TL) or exposed to green laser light (OSL). The total radiation dose the sample received during burial (termed the 'palaeodose') is determined by comparing the acquired luminescence with the luminescence produced by laboratory irradiations. The sample age is then calculated as (where 1 ka = 1000 years, and Gy is a unit of radiation dose)

Age (ka = Palaeodose (gy) + Environmental dose rate (Gy per ka).

Application of TL and optical dating to artefact-bearing sands at Malakunanja II, Malangangerr and Nauwalabila I rock-shelters in Arnhem Land revealed the great antiquity of human occup0ation in the region, suggesting human arrival 50,000 to 60,000 years ago. comparative carbon-14 and luminescence dates were obtained from the upper levels and show a close correspondence. At Malakunanja II, calibrated carbon-14 ages of 15,000 +/- 400 BP and 18,000 +/- 200 BP bracket a TL date of 16,000 +/- 3000 BP. At Nauwalabila I, a calibrated carbon-14 date of 23,000 +/- 400 BP is in sound agreement with a preliminary optical date of 26,000 + 400 BP is in sound agreement with a preliminary optical date of 26,000 +/- 5000 BP.

The lowest stone tools at Malakunanja II, were stratified within sands TL-dated to between 50,000 and 60,000 BP, with culturally sterile sands below dated back to 110,000 years. When these results were first published in Nature in 1990, criticisms of both the dating method and the associations of the dated sand with the artefacts were made by Bowdler and Peter Hiscock, but these have been answered in detail. The researchers emphasize that 'our view that humans occupied Malakunaja II by 50,000 years ago is based on a proper statistical appreciation of the uncertainties associated with each TL age and the stratigraphic constraints placed upon these uncertainties'. This view received support in 1992 when thermoluminescence dates on Nauwalabila I gave a similar age for the first artefacts there. Optical dating of the lowest levels at Nauwalabila I has revealed 'the oldest stone artefacts are securely stratified within a basal rubble unit, bracketed between 55,000 and 60,000 years ago'.

In Arnhem Land the present East, South and West Alligator rivers, together with the Wildman, Mary and Adelaide rivers to the west, all apparently once joined together and formed one large, deeply incised 'Arnhem Land River'. This flowed down a gently sloping valley enclosed between Cobourg Peninsula and Melville Island out onto the Aafura Plain. There it joined the 'Arafura River', which flowed south from New Guinea, then west into the Indian Ocean. The Arnhem Land Plateau stretches 260 kilometres from north to south, and 200 kilometres from east to west. Its edges form a steep stone escapement, rising as much as 250 metres above the alluvial plains. Large rivers flow through the escarpment in spectacular gorges. The rivers are now estuarine and full of barramundi and huge saltwater crocodiles, but estuarine conditions only came into being with the post-glacial rise in sea level about 9000 years ago. The presence of fresh water, abundant food resources and large rock-shelters for protection against the elements made this an attractive area for prehistoric settlement. It is, therefore, not surprising that several rock-shelters have been found with occupation extending back well into the Pleistocene, when they would have lain some 350 kilometres inland.

Kakadu or Gagadju is the name of the language spoken by Aboriginal people of the coastal lowlands. Aboriginal ownership of the land known as Kakadu National Park was recognised in 1976, and in 1978 the traditional owners, the Gagadju Association, leased their land to the Australian National Parks and Wildlife Service as Stage 1 of the Kakadu National Park. Archaeological work in the Kakadu region commenced in the 1960s. considerable research into the prehistory of the Top End has been done over the last three decades, largely by archaeologists from the Australian National University, such as John Mulvaney, Jack Golson and Carmel Schrire (then White), who discovered the first Pleistocene sites in the 1960s, Harry Allen and Johan Kamminga in the 1970s, and a group led by Rhys Jones in the 1980s. In the earth floors of rock-shelters a series of occupation deposits containing stone tools and charcoal from ancient campfires have been found.


The handful of excavated sites in Australia had produced some remarkable discoveries long before the new evidence of their great antiquity. Two main successive stone tool traditions were identified. The more recent was distinguished by the introduction of stone spear points about 5000 years ago. This small tool tradition succeeded the earlier Australian core tool and scraper tradition. As elsewhere, this early industry was distinguished by chunky, steep-edged flakes and cores. Horsehoof cores were found in a 20,000-year-old level of Nauwalabila, and in 10,000-year-old basal levels of Ngarradj Warde Djobkeng. The great original surprise of these Arnhem Land sites was the presence in Pleistocene levels of stone axes with grooves around their sides and a ground cutting edge. Previously ground-edge axes had only been found in Australia in contexts belonging to the last few thousand years, but since then, further specimens have been found in the basal levels of other northern sites, such as the Sandy Creek Shelter in Queensland.

Pleistocene ground-edge axes from Arnhem Land, Northern Territory. All are bifacially flaked and the cutting edge is ground to a bevel; the grooves presumably assisted the attachment of handles. Left: Axe from Malagangerr, made of hornfels, grooved on one surface (19,00-23 BP). Centre: Axe from Nawamoyn, made of porphyritic dolerite, grooved on one surface and one broken margin (21,500 BP). Right: Waisted axe from Nawamoyn, made of hornfels, and indented on both margins (21,500 BP). (After White 1967 and 1971)

The axes (figure 6.2) are generally smaller than more recent examples. They were manufactured by the 'pecking' technique, which involves fashioning them with a hammer of 'hammer dressing'. A remarkable feature is the groove or nick that several axes bear, apparently to facilitate hafting the axe-head to a handle. Until the discovery of these axes with their hafting grooves, it was generally thought that the concept of attaching stone tools to handles for greater leverage was a much later development, but now evidence has shown that the technique of hafting was definitely in use during the Pleistocene. Also, a ground-edge tool has a much more effective cutting edge than a flaked tool.  These ground-edge axes have provided some of the world's oldest evidence for shaping stone by hammer-dressing or pecking. So far, only in Japan does the technology of grinding stone to make a sharp, chisel-like cutting edge have a similar antiquity, although ground-edge tools do occur in what may be equally early Southeast Asian contexts. An age of about 15,000 years has been suggested for edge-grinding at Niah Cave in Sarawak, and edge-ground tools date back to between 27,000 and 30,000 years in Japan. These significant technological innovations of grinding and hammer-dressing thus appear to be older in the Australasian region than in Africa, Europe or the Middle East.

Two features of the Arnhem Land ground-edge artefacts support the idea that they originated locally rather than being imported. Firstly, unlike the ground-edge tools from Asia, the Australian artefacts are hatchers rather than axes. Hatchets differ from the axes by being employed with one hand instead of two, and by having a lower mass and shorter handle. The small size and light weight of the "Arnhem Land examples make it clear that they should be termed hatchers rather than axes. (However, the traditional term 'axes' has been retained in this Web site.) Secondly, the early use of hammer-dressing on ground-edge tools is confined to northern Australia. Some of Australia's oldest grindstones have also been fo9und in Arnhem Land. Three grindstones were found at Malakunanja II, associated with charcoal dated to 18,000 years. Two have flat to slightly concave grinding surfaces, and the face of one was impregnated with red and white ochre. The third was a circular grinding hollow about 10 centimetres in diameter on one of its faces. Similar grinding hollows were found in the lower levels of Nawamoyn and are common today in Arnhem Land rock-shelters. It is likely that, as well as grinding up ochre to make pigment for painting artefacts, rock-shelter walls or body decoration, grindstones and grinding hollows were used in the preparation of foods such as seeds and fruits.


Charcoal at the base of the earth floor in the rock-shelter of Malangangerr - in open woodland just west of Cahill's Crossing on the East Alligator River within Kakadu National Park - was first excavated by Schrire and radiocarbon-dated to 23,000 years old. This shelter is a deep overhang in a residual outlier of the Arnhem remains of ancient meals, are visible in the earth floor. This shell midden deposit in the upper levels dates from about 6000 to 7000 years ago until the recent past. Below it are sandy deposits, which had charcoal in their lowest levels dated to between 23,000 and 18,000 years old. The stone tools in these lower sands were characterised by flaked core tools, steep-edged scrapers, utilised flakes and, most remarkably, by small ground-edge axes with grooves on their sides. Subsequent further excavation of Malangangerr by Jones and Smith yielded a sequence going back to 32,000 years, the lowest artefacts being at a depth of 2 metres.

Malangangerr rock-shelter, Kakadu National Park, Northern Territory,
contained human occupation dating back to 32,000 years.

Malakunanja II

Malakunanja II, to the south of Malangangerr near Ngarradj Warde Djobkeng and Ja Ja billabong, is a shallow rock-shelter with faded paintings on its overhanging wall. When first excavated by Kamminga, it yielded 18,000-year-old charcoal, associated with a grinding hollow, and two flattish mortars, one bearing clear traves of ochre. Further investigation in the 1980s by Jones produced some remarkably early dates, and established Malakunanja as the oldest dated site in Australia. The layers containing the oldest human occupation were thermoluminescence (TL) dated to between 52,000 and 61,000 BP (figure 6.3 to be inserted here). Human occupation appears to have commenced abruptly above sterile sand at a depth of 2.6 metres below present ground surface, dated to 61,000 +9000/-13,000 BP (KTL162). There was dense occupation from 2.3 to 2.5 metres depth, between 45,000 +6000/-9000 BP (KTL164) and 52,000 +7,000-11,000 BP (KTL158), with more than 1500 artefacts in this lowest occupation layer. Basal cultural material included core-scrapers, flakes and amorphous artefacts made of silcrete, quartzite and white quartz; a grindstone; pieces of dolerite and ground haematite, chlorite and mica; together with red and yellow ochre. Whilst allowing for the possibility that artefacts were trodden into soft, sandy, older sediments by the first occupants, the researchers state that a conservative estimate places initial occupation at 52,000 years. Below this was a 20-metre deep, sterile sand sheet, completely devoid of any artefacts, deposited steadily from around 110,000 years ago. 

Position of thermoluminescence (TL) samples relative to the lowest occupation at Malakunanja II. Depths are metres below the surface.
The vertical spread of the TL samples is indicated. Stippling shows levels with chipped stone artefacts and other occupation debris. Horizontal lines indicate boundaries of excavation units, plotted using mean start and end depths. (After Roberts, Jones and Smith 1990c)


Some of the earliest evidence for human presence so far discovered in the Top End comes from a rock-shelter in Deaf Adder Gorge in the south of Kakadu National Park. It is an area called Nauwalabila by Aborigines, and the site is now known as Nauwalabila I. (It has also been called the Linder site after local explorer, Dave Lindner.) The rock-shelter was formed by a massive boulder which toppled off the sandy floor, on which lay stone tools such as spear points. The archaeological potential of the site was recognised by Rhys Jones as Betty Meehan when they were taken there by Dave Lindner in 1972, and excavations were carried out there by Kamminga in 1973 and Jones in 1981. Two elderly Aboriginal men, who remembered camping in the shelter as small boys with their families and thus depositing the topmost layer of habitation debris, took part in the latter excavation and helped to uncover the basal layers.

The excavation of Nauwalabila rock-shelter, Kakadu National Park.
(Courtesy of R. Jones and Australian national Parks and Wildlife Service)

Stone tools were found down to a depth of almost 3 metres below present ground level, and a radiocarbon date of about 20,000 years was obtained on charcoal at a depth between 1.7 and 1.9 metres. The top 2.4 metres of deposit was fine sand which built up steadily over a period of at least 25,000 to 30,000 years. Below the sand a layer of weathered sandstone rubble some 40 centimetres thick rested on large rocks and red sand. Stone tools lay within the basal rubble - heavy weathered quartzite flakes, which, when snapped, revealed a thick chemically weathered skin on their outer surface. Evidence of edge-grinding occurs at Nauwalabila I, where flakes of dolerite showing ground facets with striations from use-wear were found throughout the top 1.4 metres of the deposit, the lowest example of these flakes being from ground-edge axes coming from the 14,000-year-old level. Below this were pieces of volcanic rock (dolerite), which because of its hardness is the characteristic material used to make ground-edge artefacts, but these had been so heavily chemically weathered that their original form was not readily discernible. At a depth equivalent to an age of about 25,000 to 30,000 years were several extremely weathered lenticular objects of volcanic rock, which were of exactly the same shape as Schrire's Pleistocene hatchet heads, but they were too weathered for positive identification. Throughout the same 25,000- to 30,000-year-old layer were slabs of sandstone showing evidence of grinding and hard, heavy pieces of haematite (a high grade ferric iron ore) with grinding facets on their surfaces, indicating that they had been used as a source of pigment. Ochire in a wide variety of colours occurs in the lower levels of many of these Pleistocene sites, and there is some evidence that the ochre was ground up and pulverised. Small circular hollows, similar to those that are visible nowadays on flat shelves of rock in innumerable Top End rock-shelters, are clearly as old as the covering sands (i.e. more than 18,000 years) in the excavated rock-shelters of Malangangerr and Nawamoyn.

Optical (OSL) dating of the lowest levels of Nauwalabila I has shown that the earliest stone artefacts are securely stratified within a basal rubble layer bracketed between 55,000 and 60,000 years ago. As at Malakunanja II, optical dating is consistent with radiocarbon dating and with depth. The Nauwalabila site is of particular significance for several reasons. It reveals continuous occupation throughout the height of the last glaciation, when there appears to have been a break in occupation of some other sites in the Kakadu region. The antiquity of the stone artefacts in the basal rubble makes Nauwalabila possibly the oldest human occupation site yet discovered in Australia. The site also contains vital evidence of change in the landscape. When hunters first inhabited the rock-shelter, the surface of the ground was 3 metres below its present level. Since that time the ground level both within the shelter and of the extensive sandsheet plain outside has risen 3 metres. Why, straight after human arrival in the valley, was there this sudden build-up of sand, with the rate of sediment accumulation increasing a thousandfold? the answer seems to lie in Aboriginal impact on the environment. When Aboriginal people reached the Top End, they used fire extensively both for cooking and as an aid in hunting and gathering, as their modern descendants still do. In the Kakadu environment, this caused massive slope instability and erosion, leading to the build-up of sediment on valley floors such as Deaf Adder Gorge. Dramatic change came to Kakadu during the last ice age not because of climatic change, but because of the one new factor in the equation, humans with their fire-sticks.


This handful of excavated rock-shelters at the foot of the western Arnhem Land escarpment has given us the world's oldest ground-edge hammer-dressed axes, Australia's oldest grandstones and paint palette, and the earliest human occupation yet found in Australia. Thermoluminescence (TL) dates on artefact-bearing quartz sands in two rock-shelter deposits suggest the arrival of people in northern Australia between 50,000 and 60,000 years ago. The TL dates were obtained from sandy footslope deposits at occupation sites, which contain stone artefacts in their primary depositional setting. Comparative radiocarbon and luminescence dates obtained from the upper occupation layers of the rock-shelters Malakunanjua II and Nauwalabila I show close correspondence, and the validity of the dating, supported as it is by much corroborative evidence, is now accepted by most archaeologists.

The coincidence of the timing of earliest human occupation at Nauwalabila and Malakunanja, and the complete absence of cultural remains in the sterile sands in the lowest part of the sites strengthen the case that modern humans first arrived in this region 55,000 to 60,000 years ago, or, at the most conservative guess, 50,000 years ago. In other words, the fact that these desirable, accessible pieces of prehistoric real estate were uninhabited until about 55,000 years ago mans that humans did not arrive in the Kakadu region until then. This is probably the closest we will get to seeing that first footprint on an Australian beach.

Australia - Aboriginal Sites

Australian Aboriginal Anthropology

Australian Aboriginal Music

Australia Postcards and Picture Galleries
Australia Home Page
Pacific Islands Radio
Please enter your email address for your free Pacific Islands Radio Newsletter!


Hosted By Topica















Jane's Oceania Home Page Newsletter
to get the latest news, information and Web site updates!
Please enter your email address below,
then click the 'Join' button
for your free Newsletter!
 Join newsletter! 
Search Jane's Oceania Travel Page Search the Internet

Oceania Time Zones
Oceania Weather

Back to Jane Resture's Oceania Page

Back to Jane's Oceania Home Page

(E-mail: -- Rev. 21st September 2015)
eXTReMe Tracker
eXTReMe Tracker