by Tim Flannery
The first of those consequences, I hypothesised, was that as most large herbivores became extinct, and the survivors were reduced to low numbers, much uneaten plant-matter accumulated, providing fuel for more frequent and intense fires. This changed fire regime in turn dramatically altered Australia’s vegetation, allowing the eucalypts and other fire-promoting species to spread from their original habitat in the regions of poorer soil until they occupied most of the continent. Their expansion came at the expense of ‘dry’ rainforest and fire-sensitive scrub, which had supported a considerable element of the megafauna, but which after their demise declined to near extinction.
These changes in vegetation led to a dramatic shift in Australia’s climate that, American climatologist Gifford Miller pointed out, occurred because less moisture is transpired into the atmosphere from eucalypts than from rainforest trees. Before the extinctions, this moisture was carried inland on prevailing winds, enhancing rainfall by up to 60 per cent.
My hypothesis caused outrage in some sectors of academe, with a fierce reaction coming from those who believed that indigenous people always lived in harmony with nature. The hypothesis was then eminently debatable, for in 1994 we had no reliable dates on when Australia’s megafauna became extinct, and only a slim grasp of when people had arrived in Australia. Furthermore, although palaeobotanists had established that a great change had indeed occurred in Australia’s vegetation, we had no idea when this momentous shift had taken place.
The controversy was made worse by a series of misleading dates (now mostly discredited) which suggested that Australia’s megafauna survived until 6000 years ago, that the first Aborigines arrived 116,000 years ago, and that the vegetation change had occurred as recently as 38,000 years ago—or perhaps as long as 120,000 years ago. This absence of firm dates left me determined to pursue and test my hypothesis by dating the megafaunal extinction event. Fortunately, new techniques were becoming available that would allow us to reach back in time as never before.
In pursuit of this goal I found myself loitering, one chilly morning in July 1997, beside a small crevice located deep in the Margaret River district in Australia’s southwestern corner. Beside me stood three other scientists: Gavin Prideaux, then a doctoral student but now the world authority on the short-faced kangaroos; Melbourne University’s Dr Bert Roberts, an expert practitioner of a new dating technique known as optically stimulated luminescence (OSL); and the unforgettable Professor Rhys Jones of the Australian National University, a leader in the field of archaeology and one of the most innovative thinkers Australia has ever produced. Gavin explained that the crack before us was the opening to Tight Entrance Cave, a subterranean labyrinth that contained a chamber filled with ancient bones. He had excavated thousands of samples from the cavern, including the jaw of an ancient rat-kangaroo related to the bettongs but twice the size of any living species. This new discovery, he announced with a slight smile, would be given the generic name of Virginia, but eventually he opted for Borungaboodie, meaning ‘very large ground-rat’ in an Aboriginal language.
I was keen to examine this fossil deposit first-hand, but had recently been trapped in a cave in West Papua, and tight squeezes underground filled me with horror. I tried to lower myself in, but choked as I felt the limestone walls grip me. I couldn’t do it, and backed out. Rhys Jones, who was as round as a butterball, grasped the rope and forced himself into the gap. What none of us knew then was that Rhys’s shape was partly the result of a liver swollen with chronic myeloid leukaemia, a disease that a few years later would take him from us. He had been feeling unwell for months, yet on that morning the Welshman displayed such pluck it took our breath away. Down he went into the cave without a moment’s hesitation, somehow manoeuvring his tender liver around the limestone shelves and projections, to emerge several hours later with the vigour of a cork leaving a champagne bottle. After a spontaneous round of applause he settled down to display his samples, telling us what he had seen and giving us his interpretation of the deposit.
Rhys said the place was a treasure-trove of bones, laid down by ancient streams that had coursed through the caverns. He doubted, however, whether we could accurately date the remains using OSL, a technique that Bert, Rhys and I had hoped to use to date megafaunal extinction. To succeed would be a world first, so Rhys’s assessment was disappointing news indeed.
Despite its recent application, luminescence dating (of which OSL is one branch) has venerable origins dating back to the seventeenth century and the Honourable Robert Boyle, founder of the Royal Society and the father of modern chemistry. A bachelor, Boyle lived with his sister Lady Ranelagh at her home in London. ‘He is charitable to ingeniose men that are in want,’ wrote his contemporary, John Aubrey, ‘and foreigne Chymists have had large proofe of his bountie, for he will not spare for cost to gett any rare Secret.’ Nothing fascinated Boyle as much as things that glowed, glimmered or shone.
His researches into luminosity were diverse. He made close observation of rotting fish (which sometimes glows feebly) and attempted to import from the Virginia colony a species of flea that reputedly glowed in the dark. But it was his experiments in pursuit of phosphorus that most disturbed the neighbourhood. Boyle knew that phosphorus was derived from ‘somewhat that belonged to the body of a man’, and in his first series of experiments his unfortunate manservant Bilger was required to collect the pisspots of Pall Mall and boil gallons of urine in the backyard. A disgusting black mass was all that resulted, so Boyle changed tack. Human faeces, he decided, must be the source. Bilger was then ordered to collect and bake huge tubs of the stuff, after which the long-suffering manservant searched for more congenial employment. Boyle did eventually purify phosphorus from human urine (a very poor source of the material) by using greater heat.
It was around 1663 that Boyle made his signal breakthrough in luminescence. He had begun to experiment with an object called the Virgin carbuncle’—a type of diamond reputed to glow only once, when first heated, and never thereafter. Boyle withdrew to his four-poster bed with the jewel, where he ‘brought it to some kind of glimmering light’ by ‘holding it a good while on a warm part of my naked body’. This excitation of the carbuncle by bodily warmth delighted Boyle, for he was well aware that his constitution ‘was not of the hottest’. Over three hundred years later it would also delight Bert, Rhys and me, for it would help date the demise of Australia’s giant kangaroos.
In 1948 it was established that the carbuncle’s glow came from electrons trapped in the flawed diamond’s crystal lattice. As they became excited by the warmth of Boyle’s belly button (or wherever he secreted his ‘virgin’) they escaped their traps and in the process emitted a feeble light. The knowledge that electrons can be trapped in crystal lattices, and then counted when they are released through heating or light, laid the basis for modern thermoluminescence dating.
The technique used by Bert exploits the fact that sand grains are also crystals that trap electrons within their lattices. The energy imparted to the electrons is provided by the Earth’s radiation, and once a quartz grain is buried, the ‘electron traps’ in its lattice will fill at a rate determined by the intensity of this radiation. If you know the rate of background radiation at a site (and thus how quickly the electron traps will fill), you can measure the length of time the sand grain has been buried.
The only problem I foresaw with our application of this ingenious technique was that it was based on dating grains of sand, not bones. How could we be sure that the bones we wished to date were the same age as the sand that enclosed them? This was not a trivial problem, for bones buried in swamps, riverbanks and sand dunes can be re-exposed by drought or flood (which is how we usually find them), then reburied in younger sediment, perhaps becoming mixed with younger bones and artifacts in the process. Exactly the same thing can occur in caves, where underground streams or slumping of floors allows the reburial of bones in younger sediments. Faced with several instances where it was conclusively proved that megafau
nal bones had been interred in younger sediment (including the site I had excavated at Minhamite), we decided that the only safe way to proceed was to trust dates only where the sand was found encasing an articulated skeleton or part thereof.
Our reasoning was that it is virtually impossible for a stream or flood to move an articulated set of bones from one deposit to another without disturbing them. While such a conservative approach meant discounting many sites where disarticulated remains were buried in their original context, it had the overwhelming virtue of not including sites where bones had been reburied.
One key problem was locating enough sites. Bert, Rhys and I were surprised at how few ice-age sites there were in Australia that had articulated megafaunal remains. We were unsure whether articulated bones were in Tight Entrance Cave, which is why Rhys so bravely entered it to discover that, with a single possible exception, the bones were disarticulated. We could date the sand in Tight Entrance Cave, he concluded, but a question mark would hang over the age of the bones themselves.
Fortunately, Western Australia’s Margaret River region is riddled with caves, and other caverns did contain the articulated remains of megafauna. The most important is Kudjal Yolgah, which lies nestled amid a majestic grove of karri trees just south of Tight Entrance Cave. Thankfully, it is a walk-in cave, and it contains the remains of a modest diversity of large marsupials, including an extinct wombat, a large form of the western grey kangaroo and two species of short-faced kangaroos. After documenting several such deposits, and gaining material from them suitable for dating, we turned to eastern Australia.
The Liverpool Plains of New South Wales is famed for its ice-age fossil deposits, and after returning to Sydney I received news of a discovery in this region by the owners of a property enticingly called The Mystery. It was not easy to find, but eventually I located an old wooden farmhouse with return verandahs, tucked away at the end of an ill-marked dirt road. The man who had found the bones introduced himself as John. He was in his fifties, and he worked the property with his brother and mother. Neither boy had married, and the house seemed to have changed little since their childhood, except for the collection of fossils that stood proudly on the verandah, right beside the front door. All of the familiar megafauna were there—teeth and jaws of giant wallabies, marsupial lions, and the limb-bones of the mighty diprotodon—nine species in all. John had found the fossils on the banks of the Mooki River, which flowed behind the house.
The collection was the work of years, with most finds made after floods had scoured the area, revealing fresh fossil beds. Although none of the bones appeared to be articulated I hoped that such examples might exist at the site itself. As soon as we reached the riverbank and were shown where various finds had been made, however, my disappointment grew, for the bones were from a coarse gravel that formed the banks of a river severely degraded by grazing and salination. There was little chance that any skeletons would be found in such deposits, for the ancient river that carried the gravels into place would likely break them up.
Incidentally, if you are struck by the number of eroded pastoral properties I have mentioned here, it is no coincidence. Erosion is the palaeontologist’s friend, and many is the fine deposit that I have seen disappear under the carpet of green spread by Landcare programs. I’m a great supporter of Landcare, so it leaves me feeling distinctly odd, this conflict between good land management and the pursuit of knowledge of the past. I do sometimes find myself praying for the odd monster flood, to gouge at riverbanks and wash away vegetation, so that I might find some interesting fossils.
It was with little enthusiasm that I took a dating sample. This is done by hammering a PVC tube into the fossil-bearing sediment, removing it, then wrapping it in black plastic to protect the sand grains from exposure to sunlight. The dosimeter—a phallic-shaped object—must then be inserted into the hole the sample came from to establish the ‘dose rate’ of background radiation. After repeating the procedure nearby I packed up and lugged the bulky equipment to the truck. It was a process I would repeat again and again in wide-flung areas of Australia and Papua New Guinea as we pursued our three-year-long study. The Mooki River gravels, incidentally, proved to be between 38,000 and 46,000 years old, but how much older, if any, the fossil bones were was anyone’s guess.
Bert and I realised that unless we developed a more economical method of locating suitable fossils we would never complete our work on time. Perhaps one megafaunal site in ten would yield articulated bones, meaning visiting hundreds of sites before finding twenty we could use. There had to be a short cut. Perhaps, we reasoned, if museum curators had neglected to clean the bones of megafauna in their collections, we could obtain enough sediment samples straight out of museum drawers? Museums thereafter became our major research locations, and in them we found much unwashed treasure.
The Queensland Museum yielded a cluster of articulated skeletons found a few years earlier on the Darling Downs and thankfully samples of the sediment that encased them had also been collected. And a series of skeletons with enclosing earth still adhering was also identified in the collections of the Museum of Victoria. In all we obtained samples from twenty-nine sites, but only eighteen had articulated remains.
Then came a wait of months, while tight-lipped Bert conducted his painstaking analysis. He revealed his preliminary results to no one, not even his co-workers. By the end of 2000, as the silence persisted, I was becoming fearful that our work would come to nought. After all, although we picked the most recent-looking sites there was nothing to indicate that they did not span half a million years or more. If the dates were evenly spread in time, the chance of us finding even one site from the last few millennia before megafaunal extinction was remote.
Then I received a phone call. Bert was ready to reveal his findings. This was a special day, he explained, one to be relished, for today there would be just two people on Earth—he and I—who knew when Australia’s megafauna became extinct. Kudjal Yolgah Cave in Western Australia and the site from Queensland’s Darling Downs, he said, both dated to around 46,000 years ago, give or take a few thousand years (all dates discussed here have an uncertainty of a few thousand years). No site with articulated skeletons dated to after this time, but a large number dated to just before it. The result could hardly have been better, for the two sites are located on opposite sides of the continent, and between them contained a moderate diversity of megafaunal species (rather than one or two species as many sites did). Taken as a whole, our data indicated that a simultaneous extinction event had affected a significant portion of Australia’s megafauna around 46,000 years ago.
As our studies wound down and we prepared to publish our findings in the journal Science, Bert and other colleagues were using OSL to further refine the time of arrival of people into Australia.
Devil’s Lair is one of the most important human occupation sites in Australia. Located a few kilometres north of Kudjal Yolgah, its long, uninterrupted sequence of sediments document climatic, faunal and cultural change in the region for over 63,000 years. Scientists used a battery of dating techniques, including OSL, to date the initial occupation of the shelter to around 46,000 to 48,000 years ago. The equally significant site of Australia’s oldest human burial—Lake Mungo in western New South Wales—was also redated using a variety of techniques. Previously thought to be 60,000 years old, initial human occupation there was shown to be around 45,000 to 47,000 years ago. This coincidence with the time of megafaunal extinction was concordant with my hypothesis, but the story is not finished yet. A site in Arnhem Land has a scatter of stone tools in a layer of sand dated to around 53,000 to 60,000 years ago, and a second site may be equally old. Did humans reach and settle Arnhem Land 10,000 years before spreading south? Or did the stone tools drop from higher levels into the older sediment through cracks or other disturbance? At present we simply do not know.
At the time these exciting new dates were coming in, Chris Turney, a bright young researcher from Queen’s Universi
ty, Belfast, was applying the refined ABOX method of carbon-14 dating to vegetation change in Australia. This technique involves rigorous pretreatment of the carbon samples to remove all contamination, then analysis in a mass spectrometer, enabling accurate dates of up to 60,000 years to be obtained. The focus of his research was to date the sediments that record the dramatic alteration in Australia’s vegetation, from fire-sensitive to fire-promoting species. The critical sites are in Queensland—with the most important being a crater lake on the Atherton Tablelands known as Lynch’s Crater. ABOX dating revealed that the change in vegetation there occurred around 45,000 years ago, again coinciding with our dates for megafaunal extinction and human arrival.
Science is all about testing and re-testing hypotheses, and Rod Wells of Flinders University has been rigorously testing the Future Eaters hypothesis by analysing a series of sites in South Australia. He is trying to determine whether megafaunal extinction occurred earlier in the inland than on the coast, as might be expected if climate were the cause. So far all mainland sites with articulated remains he has tested are more than 46,000 years old, with little difference in age between inland and coastal sites. As part of this project an extensive and continuous record of fauna from a cavern called Wet Cave in the Naracoorte area has now been dated. The sedimentary sequence, which lacks megafauna but contains abundant remains of grey kangaroos and other large surviving marsupials, extends back 45,000 years, establishing that megafauna has been absent from the area since that time. More intriguing discoveries were made by Rod on Kangaroo Island, where articulated skeletons of several species of megafauna have been unearthed. With preliminary information suggesting that they may be less than 46,000 years old, exciting work is ongoing at this site.
A second group of researchers led by Judith Field and Richard Fullagar is examining fossils at Cuddie Springs in western New South Wales. They believe this site indicates that the megafauna survived until around 35,000 years ago, and that it overlapped with people for at least 6000 years. But the skeletons at Cuddie are not articulated, so OSL cannot be used to date the bones. The technique has revealed, however, that the sand grains from the crucial layer in the deposit are of different ages. Perhaps a mass flow of sediment, which did not expose all the grains to sunlight, caused this. Perhaps that flood also jumbled bones and artifacts of differing antiquity in the layer, for the tooth of an ancient crocodile has been found out of place there. It might also explain the co-occurrence of grindstones (of a kind used for crushing grass seed) that elsewhere in Australia date to less than a few thousand years old, with the bones of diprotodons.