BETHULIE, ORANGE FREE STATE, SOUTH AFRICA, OCTOBER 1999
I was tired, bored, hot, thirsty, and very much wanting to go back to camp and rest my sore feet in a bowl of muddy-but-cool river water. With skin pinked by the sun and wrinkled by the incessant, sandy wind, I shuffled through the dusty heat on sore legs, arriving at a fork in the small watercourse that I had been following for more than an hour, vainly seeking paleontological gold, the spectral skulls and skeletons of the long—the very long—dead. Only patches of sedimentary rock were visible above this sandy streambed, one that eventually emptied into the Orange Free State’s Caledon River, itself one of the largest watercourses in this dry region of South Africa’s Great Karoo Desert. One fork circled roughly back in the downhill direction I had come from, leading downward through time, back into the Permian period, toward the slightly older rocks at the river’s edge that were full of Permian-aged skeletons, the remains of a large and curious land animal fauna that characterized planet Earth some 251 million years ago. The other fork headed in the opposite temporal and geographic direction—up in time toward the Triassic period, the time immediately after the greatest mass extinction in Earth’s history, the Permian extinction, also known as the Great Dying or the Mother of All Mass Extinctions.
So who was the father? I mused mirthlessly, deciding to keep heading up through time, even if it took me farther from the truck and my companions. This was no small decision, especially this late in the day. While fossils had been common on the Caledon, in beds deposited perhaps a million years before the Permian extinction, they had become ever rarer as I approached the end of the Permian. This was so different from my earlier experience in studying mass extinction, for I had come here following a decade of studying the fossil record at many K-T boundaries around the world, and at every one of those places the fossil record had been very different from here indeed. The Cretaceous fossils remained common and diverse right up to the K-T impact layer with its overlying boundary clay—and there they simply disappeared. Here in these late Permian rocks, vastly older than even the Cretaceous, it was as if the world had been slowly dying over a considerable length of time. There were many possibilities for this: Perhaps the nature of the way in which these large land animals had died and become entrained in sediment, to ultimately fossilize and rest a quarter of a billion years awaiting disinterment, had changed. Perhaps rivers had dried up, and the subsequent traps for bones disappeared as well. But perhaps the animals themselves gradually became rare, as some longer-term hand slowly but inexorably closed the windpipe of a living Earth to a near-death experience.
A particularly eye-loving fly ended that reverie, and in spite of the near absence of fossil material, I began to search again for bone, any bone, jutting from the olive-colored sedimentary rocks, striving to become an automaton, a living machine bent on seeing the visual cues of ancient bone: colors, textures, shapes that would subtly call to the prepared eye and mind. I looked up for bearings: Large buff cliffs a half mile distant were made up of sandstones and red beds of Triassic age, while the greenish rocks that peeked out of each twist in the watercourse were definitely Permian, but I was damned if directions made any sense here in the southern hemisphere; the only thing that was dependable was that the sun set back over the river. If that was Triassic up there, then somewhere ahead, and not far, had to be rock that had accumulated in swamps, lakes, ponds, but mostly in river valleys during the time of that long-ago cataclysm.
I got dustier step after step and took an occasional swig of water from my diminishing supply, sweat coursing out of my skin to be immediately swallowed by the dry air. My increased concentration was almost immediately rewarded: a few broken, eroded, but unmistakable fossils of the last Permian animals, all preserved not as articulated skeletons as they were on the Caledon’s wide stratal riverbanks but as isolated bones and teeth. Here a large scapula, probably from the most common and characteristic animal of the latest Permian, the large, cowlike Dicynodon; there a tusk of another Dicynodon; and most sensational, the broken tooth of the most fearsome carnivore of that long-ago world, a gorgonopsian, or Gorgon, as the paleontologists called them. Permian, Permian, Permian, the rocks whispered to my increasingly addled mind, the heat probably insignificant compared with that at the end of the Permian period but hot enough to make a human brain continually wander, to lose focus and capability.
Eyes down on the ground as more and more outcropping began to appear along both sides of the creek bed, I rounded a corner of the increasingly higher walled gulley and nearly died of a heart attack as three elk-sized, long-horned gemsboks startled into flight, leaping upward to scramble out of the gully, flailing legs scratching a shower of pebbles and soil out of the dirt walls as they struggled in panic to run from this bipedal stranger, spraying him with sediment in the process. Heart beating stoutly, I sat down huffing, swinging the heavy pack over my shoulder onto the ground, and rummaged for food of some sort to calm myself a bit. The short burst of adrenaline spurred by the game was dissipating, lassitude returning, only an hour till pickup and camp, an hour until that delicious first beer capped a hot day of fossil collecting. An hour. What to do for yet another hour? Idly looking down at the rocks I sat on, I absentmindedly watched the purposeful ants marching to and fro before I focused on the ants’ freeway. The ant-covered sedimentary rocks were strangely colored compared with the strata that had been present all afternoon. Not the drab olive of the Permian or the bright red of the Triassic just ahead, but an anomalous candy-cane assemblage of both.
Curious now, I stood and followed the strata to the gulley wall, to be immediately confronted by a beautifully clean rock surface, obviously scoured annually by the occasional flash flood that the Karoo experiences in its June-through-September winter. If anything, the thinly striped alternation of red and olive was even more pronounced here, about a half inch of each, beds that were clearly laminated. Such beds are known to be preserved only in the absence of life. Actually, all the day’s beds originally were just like these, but soon after their formation, ancient, Permian armies of insects, nymphs, worms, crustaceans, even the shuffling feet of the larger vertebrates visiting the shallow ponds and waterholes, where the sediment was accumulating destroyed the fine laminar bedding, churning it into a mass mixed mud of one color and almost devoid of any layering at all.
Excited now, I climbed upward, for the elevation was rising toward the high sandstone hills in front of me, and my climbing took me into younger parts of the flat-lying sedimentary strata. Only a few tens of feet above the striped rocks, themselves at least a dozen feet thick, the characteristic red mudstones with their small white limestone nodules characteristic of the Triassic were brazenly visible, and within a few minutes, an eroded skull of a small Lystrosaurus confirmed the suspicion that this was the lowermost Triassic. I walked a few tens of feet higher into the Triassic strata and entered a fossil hunter’s heaven: first tens, then hundreds, of bones, showing as a characteristic and easily seen blue-white color amid the red rocks, all from the pig-sized mammal-like reptile, Lystrosaurus, the index fossil that characterizes the earliest times after the Permian mass extinction. After the long days of searching the uppermost Permian, with its beds so barren of fossils, it was sheer joy to be amid such treasure. But these fossils told little not already known, and they were all of but a single species, rather than the more than 50 species in the Permian beds below. But it was a temptation to stay here and collect ever more wondrous fossil treasures just for the joy of it.
Discipline kicked in once again—turn around, return down into the creek, arriving in minutes once again at the striped beds. There I became busy with camera and notebook, recording thoughts, observations, and measurements. The sun had dropped into its late-afternoon position, lighting the striped strata into ever-brighter relief. I had never seen rocks like this lower in the Permian nor higher in the Triassic, but memory leafed through its files, and there were vaguely remembered beds like this at the sites called Lootsberg Pass and nearby Wapadsburg. But at those two places, this part of the sedimentary transition had been highly weathered, and since the job then and now was more about finding fossils than about noting the nature of the rocks, I had thought nothing of it. But here these rocks were cleaned, presented in their best aspect, daring me to ignore them. This creek itself was a new discovery, and its trove of fossils spoke of my being the first paleontologist to ever collect it, in all probability. So no geologist had ever stood here, to see what would soon be recognized as the best-exposed Permian–Triassic boundary in the vast Karoo Desert.
But what caused the thinly bedded rocks to form, sandwiched as they were between the normal-looking Permian rocks below and Triassic rocks above? I now knew the “stripy” beds were near what had to be the P-T boundary, and, hand lens gathered in front of face, nose to rock, I began to scan for the thin impact layer that I had so often seen in the younger Cretaceous-aged rocks, and the impact layer I absolutely believed to be present here as well. Here in 1999, I still hoped to be the first to prove my—and others’—hope for the Permian, with evidence of a large asteroid impact, evidence that would show that the Permian extinction, just like the Cretaceous mass extinction, was caused by impact as most of the geology community believed—and had believed since 1990. I wanted to be the Permian version of Walter Alvarez, and if that led to a pleasant basking in the publicity and academic honors that would come to the first to truly show that the greatest extinction was an impact extinction, so be it. I could learn the necessary modesty while pocketing the pay raise. The only nagging problem was that no matter how I and so many colleagues all over the world tried, no one had been able to find any of the telltale evidence of impact so clear and abundant in both marine and land deposits of 65 million years ago, the day that the dinosaurs were very quickly killed off. That was a round hole of evidence, and all that could be found in the Karoo and the many other P-T boundary sections were the square boxes that we tried to smash our meager evidence into.
I scanned the surface with my lens again, found nothing, and patiently began the process again. Small samples were collected across the many strata making up these beds. The sound of a distant truck horn signaled the end of the day. But I would be back, dragging my companions the next morning. This place was important. It just had to be impact. How could slow climate change, or volcanoes with lava thick or thin, have caused this greatest mass extinction in Earth’s history?
WITH THE END OF THE TWENTIETH CENTURY AND THE ARRIVAL OF THE twenty-first, ever more attention was being paid to the Permian extinction, and why not—it was the largest of all, with as many as 90 percent of all species disappearing. But how fast, which is a clue to how, began to be best appreciated with the work of paleontologists from China and the United States in extensive studies of thick Permian and Triassic limestones cropping out near Meishan, China. The geologists in China even had an advantage not present at most of the K-T sites—in China there were scattered ash layers that could be dated using large machines, and this was done on samples by MIT’s Sam Bowring. The results of this vast enterprise came out in 2001 in Science, authored by Y. Jin, Doug Erwin, Sam Bowring, and other Chinese colleagues.
The China effort combined results from five different stratigraphic sections in the Meishan locality, with sampling intervals made every 30 to 50 centimeters. A total of 333 species of marine life were ultimately found in these rocks, belonging to such varied sea creatures as corals, bivalve and brachiopod shellfish, snails, cephalopods, and trilobites, among others. Nowhere at any stratigraphic horizon at any time has so thorough a collecting effort—or so rich a fauna—been documented with such precision.
The authors did indeed find one horizon where more fossils went extinct than in any other, in the last meter of strata deposited at the very end of the Permian period, and this was like the situation at the K-T boundary sites. But unlike the K-T sites, which showed but a single level of mass extinction coincident with the impact layer, these Permian sites showed many other levels in which lesser but still significant numbers of fossils suddenly went extinct in addition to this most catastrophic level. These layers both predated and postdated the P-T boundary, which was identified based on the last occurrence of small fossils known as conodonts. It was as if there had been a series of catastrophes, one big one and many nearly as big ones.
Some years earlier, one of the coauthors, Erwin of the Smithsonian, had championed a theory that he called the Murder on the Orient Express explanation: that just as there was no single killer in the great Agatha Christie whodunit, so too was the P-T event really the end result of Earth undergoing a multitude of stresses that when combined, caused the hideous mass extinction. But in 2000 Erwin came to a new view. His work with Sam Bowring on the Chinese sections had shown that the event must have taken place in 165,000 years or less, with emphasis on the “or less.” But this is still a far cry from the interval of time that was by that point accepted for the K-T die-off—not hundreds of thousands of years but perhaps just decades.
Like so many others consumed by the mysteries of mass extinctions, Erwin searched for cause among the many threads of evidence left behind in the rock record. The various environmental conditions in the seas at the end of the Permian included widespread evidence of oceanic anoxia, or low oxygenation of seawater, in both the shallow and deep sea. The anoxia was apparently of such magnitude that many marine organisms were rather suddenly killed off, just as they are today in modern red tides. There is also evidence of global warming at the time of the extinction, and the coincidence—if that is what it was—of the Siberian lava eruptions at the same time as the mass extinction. And—the elephant in the room—there had even been the sensational, mid-1980s announcement from a Chinese group that they had discovered an iridium-rich impact layer from the highest Permian rock in these fossiliferous sections.
But science is predicated on replicability. American researchers asked for splits of the Chinese samples, and to the ultimate embarrassment of the Chinese, the highly sensitive American instruments could find no hint of excess iridium. When the dust settled, there was no indication of impact from these rocks.
How to account for the various lines of evidence that did hold up, and how could they add up to a possible, single cause—if at all? Erwin summarized the various suspects. First is the possibility that the Siberian traps introduced large volumes of gas into the atmosphere, triggering large-scale climate change and acid rain, as earlier suggested by Paul Renne and others. With new information from disparate sources, a sudden methane release into the atmosphere became a viable candidate for the killer. But in spite of no evidence to support impact, the understanding that impact could cause extinction was still on everyone’s mind. The new evidence from China argued for some sort of “quick strike.” Among potential causes of mass extinction, only asteroid impact was thought to be capable of causing such mass death in so short a time. The last sentence in the report by Jin et al. says it all:
Despite the lack of compelling evidence for extraterrestrial impact, the rapidity of the extinction and the associated environmental changes are also consistent with the involvement of a bolide impact in this most severe biotic crisis in the history of life.
Thus, in 2000, the Permian extinction looked like nothing known—it was still suspected to be some sort of impact extinction by the geological community, but one seemingly different from the K-T event: perhaps many impacts or a single large impact superimposed on some other kind of extinction mechanism. The most puzzling thing was that search as they might, none of the investigators looking at the Chinese rocks could find the well-known clues so common at the many K-T boundary sites. And then, as if the geological gods had answered the prayers of geologists beseeching them for impact evidence at the end of the Permian, in one fell swoop new results from three Permian outcrops, including the crucial one at Meishan, China, fingered impact as the culprit after all. This new evidence came from an entirely new line of geochemical study. For those yearning to find impact at all mass-extinction boundaries, a strange substance known as buckyballs seemed to come to the rescue. But in fact, what they did was light an ongoing controversy.
It was in 2001 that a new character emerged center stage with a dramatic report published in Science. The senior author was a geochemist trained at the Scripps Institution of Oceanography named Luann Becker. Her colleagues were Robert Poreda and Andrew Hunt from the University of Rochester, New York; Ted Bunch of the National Aeronautics and Space Administration’s (NASA’s) Ames Research Center at Moffett Field, California; and Michael Rampino of New York University and the Goddard Institute for Space Studies. They reported finding, in the critical latest Permian boundary layers, high levels of complex carbon molecules called buckminsterfullerenes, or buckyballs for short, with the noble (or chemically nonreactive) gases helium and argon trapped inside their cage structures. Fullerenes, which contain at least 60 carbon atoms and have a structure resembling a soccer ball or a geodesic dome, are named for Buckminster Fuller, who invented the geodesic dome.
The researchers interpreted these particular buckyballs as extraterrestrial in origin, and therefore, like iridium (which, pointedly, was not found) because the noble gases trapped inside have an unusual ratio of isotopes. For instance, terrestrial helium is mostly helium-2 and contains only a small amount of helium-3, whereas extraterrestrial helium—the kind found in these fullerenes—is mostly helium-3. According to the authors, all this star stuff could only have been brought to Earth by a comet impacting Earth at the end of the Permian period (more correctly, it ended the Permian). They found this stuff by sampling as if for carbon isotopes—by taking lots of bits of rock both above and below the boundary, carrying the pesky things though U.S. customs in the cases, and then analyzing them back in the United States.
Back in the United States, the Becker team broke down its rocks in search of the buckyballs. They are not visible to the naked eye and can be confirmed as present only by using a special kind of mass spectrometer. The results were spectacular, as results from each of the samples sections in China and Japan (but not at the third site, in Hungary) flashed onto the computer monitors attached to the various mass specs. According to the authors, the Chinese and Japanese samples were striking in being packed with evidence that Earth had been slammed by a comet (or asteroid) at the end of the Permian period. Fullerenes were found at very low concentrations above and below the boundary layer at the two sites, but they were found in unusually high concentrations at the time of the extinction.
Not only was an impact confirmed, according to the team, but also the quantitatively determined mass of buckyballs even allowed them, the authors said, to estimate the size of the comet. The researchers announced that the comet or asteroid was 6 to 12 kilometers across, or about the size of the K-T asteroid that left the huge Chicxulub crater. The scientists had arrived at this size estimate on the basis of two factors—if the body were smaller than 6 kilometers in diameter, the effects wouldn’t be seen globally, as they appear to have been; and if it were larger than 12 kilometers in diameter, there would have to be more gas-laden fullerenes distributed globally.
No one likes to be scooped. By this time, a lot of scientists had been looking for evidence of impact at P-T boundary sites for years without success. Out of the blue, a new team had hit pay dirt, and much was at stake: research money, professional advancement, but most of all, pride. Scientists are human. Of course there were very sour grapes. The results from Becker et al. were intensely scrutinized by the many who had searched without success for impact evidence at the end of the Permian period, each work weighed, each number pondered, each conclusion considered. Not surprisingly (for scientists are a naturally skeptical subspecies of humans), doubts arose, e-mails flew, and long-distance telephone charges grew.
It was the estimation of the impacter size that first made a number of the mass-extinction clan suspicious of the whole thing. Later, other doubts arose, but the impacter size jumped out at many K-T veterans when the first draft of the manuscript by Becker et al. was sent to colleagues for prepublication scrutiny. Not because the impacter size was not appropriate—on the contrary, the size was the same as that causing the later K-T catastrophe. Doubts arose because the estimated size was too perfect.
Even before publication, those asked by the Becker group to helpfully vet the manuscript were pretty much brushed aside. The publication appeared, the press had a field day, and Luann Becker thus first appeared on the scientific and public stage in dramatic fashion. She was no novice in science—prior to the Permian buckyball paper, she had published a number of papers about meteorites and their chemical compositions, and, for instance, her work on the chemical composition of some well-known meteorites in collections had nicely increased understanding of the chemical compositions of some of these widely varying kinds of rocks from space.
But it is safe to say that until 2001, she was but another of the army of scientists trying to figure out the chemistry of the cosmos. But that relative anonymity utterly disappeared with the buckyball article in Science, and soon word of this discovery was trumpeted by its major funder, NASA, in a large press conference held in Washington, D.C. NASA even invited the most experienced of all Permian workers, Doug Erwin, and he was surely bemused by all this sudden attention on a problem that he had rather anonymously worked on for years. All of a sudden the Permian extinction was one of the hottest scientific issues going—impact had once again done its magic with the press.
It would seem to be hard to follow up the 2001 circus about buckyballs and the Permian extinction with anything as dramatic, but two years later the Becker crew did just that, using the same sequence of discoveries that had characterized the history of K-T research: In 2003 they announced, again in Science, that they had found the crater of the P-T impacter itself, the source of all of those buckyballs and helium-3. Once again the whole issue made big news, the reports of two years earlier that the Permian extinction was now “proven” to have been caused by an impact quite forgotten. And this was not the only candidate “crater” to be found. It mattered not that the most experienced student of impact craters and their origin anywhere in the Solar System wrote that the structure identified as the Permian impact crater by the Becker team looked like no other impact crater in the Solar System. (Simply look at our Moon or Mars or Mercury to get a sense of how many impact craters there are in our cosmic neighborhood. To be unique among such a large number is pretty unlikely—or, the structure in question is not an impact crater).
And this was not the end of things. In 2006, a team from Ohio State University announced (at a scientific conference, not in print) that a large structure far beneath Antarctic ice was probably “the” Permian crater—but this was really bad science, since they could neither confirm that the large structure, detected remotely using gravity anomaly measurements (large craters give a different gravity reading than surrounding rock), and most important, since they could not reach any of the buried rock, they had no way of knowing what age their “crater” was. Nevertheless, once again the press grandly announced that the Permian extinction mystery was solved and that it was caused by impact.
Back in 2001, Becker and crew probably expected to be met with open arms and praise by the many scientists who had shown that the K-T event was caused by impact. But there were not a few cold shoulders. Some of this might have been jealousy, for who does not love the attention of the fickle press, especially scientists trying to get funded. Yet there was more to this doubt than that. There was a nagging unease about the data. The whole issue of buckyballs and helium-3 had yet to be accepted or even, most important, replicated by other labs. There was also a distinct feeling that not everything written by Becker et al. added up scientifically. For instance, one of the lead paragraphs of their first press release stated: “The collision wasn’t directly responsible for the extinction but rather triggered a series of events, such as massive volcanism and changes in ocean oxygen, sea level, and climate.” This conclusion made no sense at all.
How could a comet impact create volcanism or a change in sea level? Much was known about what large-body impact on Earth could or could not do, and this was in the realm of the “could not do.” While it makes intuitive sense that a large rock slamming into Earth could somehow shake free some great volcanic paroxysm, that doesn’t mean that it will. About this time many workers became less sanguine about the possibility that the P-T extinction had been caused by impact, leaving buckyballs and dead species in its wake.
The first meeting of what was to become a loyal (to science) opposition was convened only several days after the initial publication by Becker et al. in 2001, and it was by sheer coincidence that an eminent group of specialists most concerned with the P-T came together. Such a gathering was sure to eventually take place, but it might have been a year afterward or more, perhaps much more, at some scientific meeting or the other, that these same scientists would have most certainly compared notes on the purported Becker discovery, if discovery it indeed was. For other reasons than talking about the P-T extinction, Doug Erwin of the Smithsonian and Yukio Isozaki of Japan arrived simultaneously at the Division of Geological and Planetary Sciences of the California Institute of Technology in Pasadena. Already there were two California Institute of Technology (Cal Tech) faculty members also immersed in P-T research, Joe Kirschvink and Ken Farley. They decided to spend an afternoon going over the Science paper by Becker et al.
These four sat around a table, chewing on the Becker paper, and were immersed in the give-and-take of critical science. Of the four there, Farley was by training the most versed in the primary argument of the Becker report—that helium isotopic ratios could provide evidence of a past asteroid or comet collision with Earth. In addition, Farley was already acquainted with Becker, as both had been grad students at Scripps Institution of Oceanography, in La Jolla, California. Following his thesis research, Farley had gone on to do work on noble gases, and in fact by that time he was recognized as the world’s authority on helium in rocks. Among the others around the table, Erwin was a macrofossil paleontologist and the acknowledged expert on the extinction of larger fossils at the end of the Permian period, while Isozaki nicely complemented the paleontological side of things, as his specialty was the identity and fates of microfossils before, during, and after the mass extinction. Kirschvink, the last member of this group, had by this time spent several years in the Karoo of South Africa looking at the P-T boundary. It would have been hard to find a better group for critical analysis of the data from Becker et al.
As the first step in this process, the small group studied the data tables in the paper by Becker et al., fixing on the amount of buckyballs found at the three sites. While Becker had touted Hungary as yielding the crucial carbon compounds, once the four dug deeper into the data part of the article, they concluded that the Hungary site showed no evidence of fullerenes, so the critical evidence came from the other two sites. Both of these suites of rocks—from China and Japan—were intimately familiar to the assembled group at Cal Tech, for Erwin and his colleague Bowring had collected the Chinese samples analyzed by Becker, while Isozaki had done the seminal work on the Japanese P-T boundary that Becker and her group had analyzed. And it was here that Isosaki let loose his bombshell, still unknown to Becker.
Geologist Rampino, an absolutely die-hard proponent of impact and one of the authors of the Science paper, had collected the samples from Japan from seriously deformed and highly fractured deep-sea sediments. Isozaki, a specialist on deep-sea microfossils of this age, had gone back, following sampling by Rampino but prior to the publication by the Becker group, and looked at the site of the crucial samples. The places sampled were obvious, and if they were of latest Permian age, they would contain a specific set of latest Permian microsamples. To his surprise, the fossils found by Isozaki from these rocks were not the Permian species at all but were Triassic in age, from a time when absolutely no extinctions took place! Unknown to himself or any other members of the Becker team, Rampino had sampled rocks far younger than the crucial Permian age! Any results had nothing to do with the Permian mass extinction.
Thus, with no buckyballs from the Hungarian samples and the discredit of the Japanese samples, that left only the samples from China as proof of an impact. And it also left a huge residue of unease among the P-T specialists. Kirschvink, Erwin, and Isozaki turned to Farley, for he alone would be able to repeat Luann’s observations on the critical Chinese P-T samples. Farley had by that time come up with a less laborious way of detecting helium-3 from rocks. After all, it was not the buckyballs that were so important (for they can be made on Earth, in forest fires, for instance) but the fact that they encased the helium-3. Farley bypassed the buckyball question entirely, going straight after the amount of helium in the rocks. He even tested his new method—on K-T samples that Kirschvink and I had obtained in Tunisia the year before—and had found helium-3 in them, as was expected. Now, at the urging of all, Farley turned his attention to the Chinese samples.
All science is predicated on replicability, as belabored earlier in this chapter for a reason. Farley contacted Becker, asking for splits of her critical Chinese samples. She replied that she had used up all of her samples in the analysis and could supply no more. This was curious and unsettling—how could she have not saved some of the samples so that others could do just what Farley was trying to do—replicate her results? Farley then went to the original source of the samples, MIT geologist Bowring, who had actually taken the critical samples from the Meishan locality that was now so crucial for understanding the ancient mass extinctions. Bowring promptly sent new material from China, which was duly analyzed. Farley used a blind sampling technique, asking Bowring to withhold any information about which samples came from the critical level where Becker had found the helium-bearing buckyballs. After exhaustive tests, Farley was not able to replicate the Becker group’s findings. There was no helium-3 to be found in the Chinese samples examined by the Cal Tech lab specializing in this kind of work.
Speculations were rampant after this surprising development. Becker was told of these results and shrugged them off; the most likely reason for this negative finding, she reasoned, was that the helium-3 layer discovered by her group was an extremely thin layer from a more massive sample collected and supplied by Bowring and that the material later sent to the Cal Tech group did not contain this exact bit of rock. But others came to other conclusions. There was some speculation that the Becker findings might have somehow been related to lab error or contamination of Becker’s glasswear in some fashion.
Becker pressed on, of course, and as recounted above, continued in the lines of K-T science by following up the geochemical discoveries with an announcement that the team had found the crater left behind by the comet spewing all those buckyballs over the planet (but bypassing Hungary, anyway) in the paper published in 2003. The Bedout crater, as it was named, lay underwater off Australia and was certainly large enough, containing rocks from within that were of approximately the right age. Another rock from space (or ice ball, in this case) did the damage. Impacts cause extinctions, a paradigm again verified.
Soon thereafter, letters bombarded Science, demanding to know how such work could get published, and Science went silently defensive (probably to the delight of the competing European journal Nature, for some of the harshest critics of the whole Permian impact story were European impact specialists). But as far as public relations went, who cared? It was a good story, tidily completed.
Thus, by the middle of the first decade of the new century, the riddle of the cause of the Permian extinction was solved, at least in the press’s and public’s minds, by the discovery of nonreplicable helium-3 findings from a noncrater crater. What a disconnect between the public and the on-the-ground scientists!
So if not impact—what? In the first five years of the new century, two camps emerged, each deeply entrenched in its views: Either the Permian extinction was caused by impact or its cause was unknown but certainly not impact. In favor of the former was the discovery by the Jin and Erwin group of a sudden extinction exhibited by the fossil invertebrates. How else but impact could this have occurred? Yet many lines of evidence were converging on something more prolonged than a single quick strike. In September 2000, University of Oregon geologists Evelyn Krull and Greg Retallack published a paper detailing their results from prolonged geological and geochemical studies of P-T boundary sections in Antarctica. Their results strongly supported the idea that the early Triassic was a time of heightened methane gas volumes in the atmosphere. Methane is one of the most potent of the greenhouse gases—and its sudden release would have driven global temperatures sharply higher. These results followed on Retallack’s 1999 findings from the Sydney Basin in Australia. There, Retallack recognized that the P-T boundary was coincident with the formation of the last coals anywhere on Earth for many millions of years of Triassic time.
The boundary coincided with a large-scale extinction among plant species as well as a dramatic changeover in climate, as deduced from fossil flora and fossil soils. A deciduous flora adapted for a humid but cold temperate climate characterized the latest Permian of Australia. At that time, Australia, like nearby South Africa, was located far nearer the poles than the equator. In the earliest Triassic, however, a marked change in climate apparently occurred. The fossil soil types indicate a much warmer climate—as would occur from a sudden onset of global warming. Coal formation suddenly ceased. Sedimentation rates markedly increased in the lower Triassic rocks, and Retallack interpreted this as being the result of extensive and sudden deforestation at the P-T boundary.
Other suggestions of a profound world-changing event came from Roger Buick, a geoscientist from Australia. Buick, a specialist on the Precambrian world, became intrigued with the P-T event because of how it sent our world, for a short time, back to conditions quite like those prior to the rise of complex animals and plants. None of the observed evidence suggested a single asteroid impact. Buick described the event in Australia as follows:
Clearly, a single impact could not have been responsible. The most obvious interpretations are repeated environmental perturbations, such as methane hydrate melting pulses, repetitive overturn of a stratified ocean and/or persistent prodigious volcanic exhalations, or serial extra-terrestrial insults. Resolving which of these is the most viable explanation for the range of geological, biological and biogeochemical features occurring over the extinction period is the aim of future research.
Serial extraterrestrial insults? Not even the Becker camp was arguing that more than a single rock from space was involved.
So where does this leave the impact hypothesis for the Permian? While popular science magazines such as Discover still promote the press-friendly impact hypothesis for the cause of the Permian extinction, among working scientists this is a rejected hypothesis. Once again, however, the all-too-common disconnect between what the majority of scientists believed and what a few media-savvy scientists believed led to very different points of view about the Permian extinction. The impact explanation continued to have support because of brilliant public relations work by the Becker team. Sooner or later, however, there would have to be new data if the conflict in opinion was to be resolved.
To settle the impact question, the Permian community took a page from the K-T days—use a neutral scientist as a referee to oversee the collecting of samples by proponents of both sides, and then have the referee randomize and distribute those samples to be examined, so that each side was testing some they had collected and some they had not, without knowing anything specific about their provenance or distance from the boundary between the two periods. Funded by NASA, a group that included Becker, Erwin, and impact specialist Frank Kyte of the University of California, Los Angeles (UCLA) as the neutral “referee” traveled in 2004 to the famous Chinese outcrop itself to see if the Becker results could be replicated. Small chips taken from the sampled rocks eventually were sent to various labs across the United States, including Becker’s, but before the various labs could begin to analyze the geochemistry of these rocks, Becker and her crew quit the program, protesting that the wrong rocks in China had been sampled, even though she was there as they were sampled. The labs that completed their work found no evidence of buckyballs, no evidence of an impact.
THE FIRST EVIDENCE POINTING TO A PROCESS VERY AKIN TO THE REAL cause of the P-T mass extinction had by this time been known for nearly a decade. In 1996 a group led by Harvard paleobotanist Andrew Knoll published a startling new theory to account for the mass extinction, built on a realization that the end of the Permian period was much like the end of an earlier era, the Precambrian era, the time about 600 million years ago immediately preceding the advent of large animals and skeletons, and for much of his career, the focus of Knoll’s research.
No one else was in a position to recognize a similarity between the two. Geologists typically concentrate their efforts on one narrow time, and this similarity came to light only when Knoll decided to jump to the Permian, taking his knowledge of the Precambrian time with him. Like the Permian, the Precambrian ended with large-scale swing in the ratios of carbon isotopes in the atmosphere and a mass extinction. Knoll and his colleagues argued that the cause of both was the same, and in their paper they proposed a novel explanation for how the changes transpired.
Knoll et al. suggested that the oceans of the late Precambrian era and the late Permian period were unlike those we have today—they were stratified, with water with more oxygen on top, and less below. Furthermore, these strange Permian oceans had large amounts of organic material locked in bottom sediments. Then, for reasons still unknown (but probably related to an increase in plate tectonic activity as well as a change in the continental positions), this pattern changed. The ocean somehow changed its state so that the deepwater, which had been safely locked away from the surface for so long, began to liberate its load of dissolved carbon, in the form of vast quantities of carbon and organic material, back into the shallow waters of the sea, and ultimately into the atmosphere, as large bubbles belched forth as though the oceans were a large soft drink. At the same time, one of the greatest episodes of volcanism known in Earth’s history took place in Siberia, releasing more carbon dioxide directly into the atmosphere. The mechanism touted by the authors was akin to the horrific catastrophe that occurred in the 1980s at Lake Cameroon, in Niger, Africa. While thousands of humans and their livestock slept, the deep volcanic lake burped to the surface and into the air a gigantic bubble of carbon dioxide. This bubble spread out over the shoreline, killing most humans and animals there, before finally dispersing into higher altitudes, driven by winds. Was this the same mechanism that happened at the end of the Permian period, only writ much larger? Were all the oceans suddenly burping up bubbles of deadly carbon dioxide and other volcanic-like gases, such as methane?
Knoll and his colleagues proposed that the sudden increase in carbon dioxide, dissolved in the ocean, killed most marine species. Carbon dioxide in elevated concentrations is a known killer, and marine animals—especially those secreting calcareous shells—are particularly susceptible to carbon dioxide poisoning. The problem with this model, however, is that it cannot explain the coincident killing of land animals. Most terrestrial creatures are less sensitive to excess carbon dioxide.
There was a great deal of discussion both pro and con following this article. While Knoll and his colleague Richard Bambach began looking at various terrestrial organisms to see how susceptible they are to carbon dioxide poisoning, oceanographers, looking at the physics required to liberate large bubbles of carbon dioxide out of ocean water, could not get their computer models to corroborate that this event could take place. No oceanic carbon dioxide, no Permian event. No one was saying that there was not copious carbon dioxide in the atmosphere—only that the carbon dioxide could not have been released quickly enough to kill anything.
Knoll’s idea—that it was gas released from the deep ocean that was the cause of the Permian extinction—lay fallow for nearly a decade. But a variant of the same mechanism came forward in 2005, and it does present a plausible mechanism for compounds held in the sea to poison life on land. It was stimulated in no small way by the fossil record of land animals across the Permian–Triassic boundary. Much of these data were mine.
THE ROCKS AND FOSSILS THAT HAD FORMED THE BASIS, PRO AND CON, of the discussion of a Permian impact had all been deposited in the sea. But evidence for what was happening on land began emerging right about the time that the discovery of the alleged Bedout crater was announced in 2003. The evidence for methane release separately published by the Retallack team (for rocks in both Australia and Antarctica) and the Buick team (rocks in Australia) were both from strata that had originated in terrestrial settings, and new studies carried out in Greenland derived their evidence from fossil plants. Other evidence of fossil plant records appeared based on new studies in Greenland. And some of the most interesting new evidence of what happened to land animals came from studies that I had by then been conducting for nearly a decade in the Karoo Desert of South Africa, the most prolific fossil boneyard of late Permian and early Triassic age in the world.
For decades a succession of paleontologists has trekked into the vast wasteland of the Karoo to retrieve ancient bones. Early on, the fossils were collected with little regard for where they were found geographically, and with even less regard for their precise stratigraphic level. But since the 1980s, a new generation of paleontologists, led by Roger Smith of the South African Museum and Bruce Rubidge of the University of the Witwatersrand, have brought rigor to the field. Their most recent work focuses on collecting just below and above the P-T boundary to test ideas about the severity and abruptness of the extinction among larger land vertebrates of the Permian period. Other, more novel approaches were brought to the Karoo by us American paleontologists who partnered with the South Africans to apply the kinds of isotopic studies that had been so successful at other geological boundaries, such as that at the end of the Paleocene epoch.
Two important findings emerged from this work: First, the isotope record showed not one perturbation but several. Second, while there was one interval when many species went extinct, it seemed to have lasted at least several thousand years, and there appeared to be smaller-scale extinctions both prior to and soon after this.
Thus, the land fossils seemed to show the same pattern as the marine fossils: a series of antibiotic events, the “insults” so colorfully described by Roger Buick. Maybe one of these was caused by impact, but by this time, teams from the University of the Witwatersrand, in combination with Christian Koeberl, another veteran of the K-T wars and an expert on impact evidence, could find no evidence of an impact at the levels in the Karoo where the highest rate of extinction seemed to occur. It’s negative evidence to be sure—and no one to date has looked for buckyballs in this section—but nonetheless certainly not evidence of a single K-T–like asteroid strike.
THE WORK OF MANY INDEPENDENT TEAMS EVENTUALLY REVEALED A PICTURE of the P-T boundary that seemed to show a succession of death intervals spanning a few million years before and after the event that marked the boundary. The picture was incomplete, however, because another line of evidence that had been collected from virtually every P-T section had been largely overlooked. These were the carbon isotope studies.
At the K-T boundary, the isotopic evidence had demonstrated that a lush, plant-filled world went suddenly dead, remaining so for tens to hundreds of thousands of years. During that period, with so many plants and photosynthesizing marine plants having been killed off by the environmental aftereffects of the impact, the carbon cycle suddenly had a glut of carbon-12 that in happier days had been tied up in plants. But that was as far as this signal went: rapid catastrophe, followed by repair of the ecosystems and replacement of the killed-off species and individuals by the newly evolved and newly grown. Pretty quickly, things were back (at least from the carbon isotopes’ point of view) to where they had been before the impact.
But a funny thing became apparent when similar kinds of studies were conducted on late Permian and early Triassic rocks. At the P-T boundary there was indeed a perturbation indicating that plants rapidly died off. This was no shock—the fossil record had already convinced everyone that many plants had gone extinct. But that was not the end of it. Unlike the K-T, where the disruption of the normal isotope record was pretty rapidly healed following the one blow, the P-T record showed a succession of perturbations. If the world at the K-T were a boxer, it would be one caught unaware by a hard cross, knocked down but soon back on his feet—not overmatched, just surprised. The world at the P-T, however, was like a featherweight fighting Joe Frazier. Knocked down, it got up, just to be knocked down again. And again and again, as our foolish world kept evolving new plants (and animals, although it is the fossil record, not the isotopes, that tell you this), just to get its block knocked off again by whatever nasty boxer the P-T extinction mechanism (mechanisms?) really was. Now this was a surprise, and one that was hard to explain.
In a movie, at this point, some new brilliant scientists muttering abracadabra, among other mumbo jumbo, would appear on-screen, pull a slick new machine out of his or her sleeve (I would love to see Lara Croft take on the Permian), and solve the problem. “The Permian extinction is solved: It was caused by…caused by…”
But this is not the movies. By 2004 we were just beginning to find out what did not do it. To better understand the Permian extinction, still other mass extinctions had to be studied.