The Great Tragedy of Science

In 2007, the hottest new idea was that an impact hit North America 12,900 years ago and wiped out the Ice Age megamammals. How has that hypothesis fared in the past 4 years?

Mass extinction is box office, a darling of the popular press, the subject of cover stories and television documentaries, many books, even a rock song…At the end of 1989, the Associated Press designated mass extinction as one of the “Top 10 Scientific Advances of the Decade.” Everybody has weighed in, from the economist to National Geographic.

—David Raup, 1991

For every problem, there is a solution that is simple, neat, and wrong.

—H.L. Mencken

The great tragedy of Science—the slaying of a beautiful hypothesis by an ugly fact.

—Thomas Henry Huxley

I vividly remember running into my good friend, Jim Kennett (now retired from the University of California, Santa Barbara), at the 2007 meeting of the Geological Society of America. Kennett is still one of the giants and pioneers of the fields of marine geology and paleoceanography and climate change, with a career that goes back to the early 1970s when the Deep Sea Drilling Project began to revolutionize our understanding of oceans and climate. As a co-author on the paper, Jim was excited about this hot new idea that an impact had struck about 12,900 years ago and was responsible for the extinction of the Ice Age mega-mammals—one of the most interesting and controversial events in earth history. I tried to sound enthusiastic, but I’d seen many versions of the impact hypotheses for other mass extinctions crash and burn, so I didn’t want to pronounce judgment yet.

This idea is the most recent entry in the scientific bandwagon that impacts caused all mass extinctions. Firestone et al. (2007) claimed that the extinction of the Ice Age “megamammals” (large mammals over 40 kg in weight) was due to the impact of an extraterrestrial object about 12,900 years ago. Naturally, when this idea was first proposed, the media had a field day, and almost no dissenters or critics were heard at all. Some geology textbooks even inserted this untested idea into their new editions without waiting to see if it would pan out or not. And just like every other half-baked idea from the impact advocates, the “late Pleistocene impact” scenario has been shot down by a whole range of observations.

The late Pleistocene impact hypothesis was born from observations that there was a distinctive “black mat” organic layer in several localities across the southwestern U.S., immediately above the last appearance of Ice Age megamammal fossils. These include not only the huge mammoths and mastodonts, but also ground sloths, horses, camels, two genera of peccaries, giant beavers, plus predators such as short-faced bears, dire wolves, and sabertoothed cats—but not bison, deer, pronghorns, and a number of other large mammals still found in North America today. The “black mat” is also above the first known artifacts of the Clovis culture, which were thought to be the first human arrivals from Eurasia, and allegedly responsible for overhunting the megamammals to extinction. Firestone et al. (2007) also claimed to have found “nanodiamonds”, iridium, helium-3, “buckyballs,” and a number of other geochemical and mineralogical “impact indicators” in the “black mat” layer, and then painted a variety of different (and conflicting) scenarios about the impacting object (they are not consistent as to whether it is a comet or an asteroid) hitting near the Carolina Bays region. This supposedly affected the Laurentide ice sheet in the northeastern part of North America and triggered the Younger Dryas cooling event at 12,900 years ago.

The entire scenario has been completely demolished by a number of lines of evidence. As Pinter and Ishman (2008) showed, there is no evidence that there was an impact in the Carolina Bays, and most of the alleged “impact evidence” is questionable when analyzed by other labs. Firestone et al. (2007) argued that the impact was an airburst, since there is no crater, no tektites, no shocked quartz or other high-pressure minerals, which are the best indicators of a true impact. Most of the material that was allegedly impact derived (nanodiamonds, iridium, helium-3, “buckyballs”, and so on) is also consistent with the normal rain of micrometeorites, and not abundant enough to be good evidence of an impact.

The claim that the “black mat” was an impact layer has also been debunked. It is more likely an indicator of a high water table and wetter conditions associated with the abrupt Younger Dryas cooling event (Haynes, 2008). The supposed “instantaneous” extinction of megamammals at this horizon has also been debunked, since the extinctions were scattered across a wide geographic area with different genera going out locally at different times (Grayson and Meltzer, 2003; Fiedel, 2008; Scott, 2010). Mammoths, mastodons, giant deer (“Irish elk”), ground sloths, and many other megamammals did not die out at 12,900 years ago, but survived in most cases to 10,000 to 11,000 years ago. This is fatal to the idea that a single impact killed them all off. In fact, none of the well-dated extinctions occur at 12,900 years ago. Most of the extinctions are either significantly younger than that interval, or there are no good final dates for their last appearance—but very little appears to happen to the megamammals at precisely 12,900 years ago.

Particularly striking is the persistence of mammoths and ground sloths well into the Holocene (as young as only 6000 years ago), and of course, the bison, deer, grizzly bear, cougars, peccaries, and pronghorns that are still with us, while elk and moose came to North America at this time (rather than being wiped out). In fact, studies of DNA trapped in soils from the Canadian Arctic shows that many of these “extinct” Ice Age mammals persisted well into the Holocene, even though there are no bones preserved in beds that young. The impact hypothesis does nothing to explain the selectivity of this extinction. In addition, the South American, Australian, and Eurasian-African megafaunal extinctions are not synchronous with the alleged “impact,” so it does nothing to explain their demise.

The claim that the “impact” had a severe effect on human cultures has been completely shot down as well (Buchanan et al., 2008), since there is no evidence whatsoever that human cultures changed dramatically at this time, or that there was a major population decline. Clovis culture was gradually transformed into Folsom, Dalton, and Eastern U.S. Paleoindian cultures, and they apparently spread widely at this time, rather than declining. And just before my 2011 Catastrophes! book came out, Jacquelyn Gill of the University of Wisconsin presented a paper at the Ecological Society of America meeting analyzing the details of lake sediments from the northeast, which preserve a high-fidelity record of that time. She found no evidence of the impact debris that was supposed to be common—and her data were gathered even closer to the alleged impact site than the evidence garnered from the western U.S. Nor was there any great shift in vegetation, pollen, spores, or any other biotic signal that would be consistent with the impact hypothesis.

Finally, if the authors of the Pleistocene impact scenario had paid any attention to the past decade of research on impacts and extinctions, they would have realized that the “impacts cause extinction” notion is passé. As I discussed in Chapter 11 of my new book Catastrophes!, none of the great extinctions of past (except possibly the end-Cretaceous event) are associated with impacts.  It feels like the Firestone et al. (2007) impact scenario is a bad rehash of the debates from the 1980s. Apparently, the authors are still stuck on a bandwagon that has long since ground to a halt—except in the popular media. As Barnosky et al. (2004) showed, the causes of the late Pleistocene megafaunal extinctions are complicated, and probably involve a combination of both human overhunting and climatic change. One thing that doesn’t seem to be relevant is an impact.

Like many other trendy ideas in science, it made a big splash when it first came out in 2007, and some textbooks even jumped the gun and featured it in their new editions. But eventually the scientific review process works through all the hot ideas that have made it past the first level of peer reviews. After 2-3 years, the majority of these faddish proposals die a quiet death as they are debunked, one claim after another. Yet the public and press only remember the splashy coverage when the idea was first proposed, and don’t realize that it has been quietly discredited in the scientific community.


  • Barnosky, A.D., P.L. Koch, R.S. Feranec, S.L. Wing, and A.B. Shabel. 2004. Assessing the causes of late Pleistocene extinctions on the continents. Science 306: 70–75.
  • Buchanan, B., M. Collard, and K. Edinborough. 2008. Paleoindian demography and the extraterrestrial impact hypothesis. Proceedings of the National Academy of Sciences 105:11651–11654.
  • Fiedel, S. 2009. Sudden deaths: the chronology of terminal Pleistocene megafaunal extinction, in Haynes, G. (Ed.), American Megafaunal Extinctions at the End of the Pleistocene. New York: Springer, pp. 21–38.
  • Firestone, R.B., and 25 others. 2007. Evidence for an extraterrestrial impact 12,000 years ago that contributed to the megafaunal extinctions and the Younger Dryas cooling. Proceedings of the National Academy of Sciences 104, 16016-16021.
  • Grayson, D.K., and Meltzer, D.J. 2003. A requiem for North American overkill. Journal of Archeological Science 30:585–593.
  • Haynes, G. 2009. Estimates of Clovis-era megafaunal populations and their extinction risk, in Haynes, G. (Ed.), American Megafaunal Extinctions at the End of the Pleistocene. New York: Springer, pp. 39–54.
  • Pinter, N., and Ishman, S.E. (2008) Impacts, mega-tsunami, and other extraordinary claims. GSA Today, 18(1):37–38.
  • Scott, E. 2010. Extinctions, scenarios, and assumptions: changes in latest Pleistocene herbivore abundance and distribution in western North America. Quaternary International 217:225–239.

rhinoceros giants

The huge hornless Asiatic rhinos known as indricotheres were the largest land mammals that ever lived, wandering from Mongolia to Turkey across dry scrublands from 34-23 million years ago. Their sheer size poses many questions about how they lived, yet we can also make some educated guesses about their ecology based on the constraints on living mammals.

When I started my graduate career at the American Museum of Natural History in New York in 1976, I soon realized that I had stumbled upon an incredible opportunity. In addition to the world-famous fossil halls that have amazed generations of visitors, there are at least a hundred times as many fossils stored in research collections for study by qualified scientists. This is where the real work of paleontology takes place: specialists dedicated to the study of one group of organisms spending weeks to months to years examining every fossil in the collection, trying to reconstruct their anatomy, determine their relationships, and decipher what is the correct taxonomic name for any group of specimens. Without this fundamental work determining which species are valid, and when and where they lived, all other work in paleontology (especially computer models which are based on counting taxa studied by others and compiling them into databases) is “garbage in, garbage out.”

The American Museum is particularly important for such research, because it has the original collections of pioneering paleontologist Edward Drinker Cope, collected from the 1870s and 1880s, plus the huge numbers of fossils accumulated by its legendary paleontologists from 1895-1935 (Henry Fairfield Osborn, William Diller Matthew, Walter Granger, and others), as well as later collections obtained by the most brilliant paleontologist of the twentieth century, George Gaylord Simpson. The collections of dinosaurs, other reptiles, birds, amphibians, fish, and other vertebrates huge, but they are all outstripped by the gigantic collection of (mainly North American) fossil mammals. In the 1920s, the millionaire Childs Frick (son of the robber baron Henry Clay Frick, Andrew Carnegie’s partner) became interested in the origin of the mammals he used to shoot on big-game hunts. Starting about 1930 and for the next 35 years he used his wealth to pay for field crews to work year-round in the important fossil beds of the western United States, making giant collections from key localities and finding many more localities. Consequently, where we used to have just isolated teeth and jaws and maybe a skull of most mammals, the Frick Collection usually has many complete skulls or skeletons. This allows a paleontologist to see the complete anatomy of a particular mammal, examine variability within a population, and determine a much more informed and modern classification of names that had been based on isolated scraps of teeth described a century ago. Thus, most of the major groups of North American fossil mammals have to be completely restudied using the huge Frick Collection before we can make any conclusions about how many species existed, and when and where they lived.

To give you a sense of the size of the collection, there is a separate wing to house just the fossil mammals, built in an interior courtyard so it is invisible to the public. The fundraising and construction was started after Frick died in 1965, and not completed until shortly before I arrived in 1976. The Frick Wing has 10 floors altogether: an entire floor of rhinos, an entire floor of camels, an entire floor of horses, an entire floor of mastodonts and mammoths, three more floors of other groups of mammals, and the top three floors are the prep lab, the offices, classrooms, library, teaching collections, and other essential spaces. The horse floor is largely studied and published, but almost nothing has been done on the camel floor or the mastodont floor. When I arrived in 1976, the Museum’s curatorial assistant Dr. Earl Manning took me under his wing and introduced me to the study of North American rhinos, which had been neglected since the 1920s. At first we worked on projects together, but after he left in 1980, I continued working on the rhinos for another 25 years, finally publishing my comprehensive book-length monograph on them in 2005. Where once rhinos were a mess of invalid species, outdated names, mistaken identifications, and uncertain relationships, now they are one of the best-documented groups of North American mammals. Using my book, you can identify any bone of any North America rhino to genus and species.

In addition to the North American rhinos, I spent a lot of time looking at the gigantic specimens of the huge hornless Asian rhinos known as indricotheres. The American Museum has the best collection of them outside Beijing and Moscow, acquired by the legendary Central Asiatic Expeditions to Mongolia in the 1920s. After my 2005 North American rhino book, I thought it might be fun to write about these amazing creatures, which weighed as much as 20 metric tonnes, larger than the largest elephant or mammoth. When I had a sabbatical in late 2011, I finally had a chance to sit down and write a book about them, and the book has just appeared.

The first thing to realize is that much about what you see about extinct animals on TV documentaries is artistic guesswork, not based on any hard evidence. In the case of indricotheres, we have only the bones, and only partial skeletons at that. There is no direct information about the color of the animal, skin texture, what it ate, how it walked or how it behaved or sounded. All of this information, so often a crucial part of the CG animations that now dominate most documentaries about prehistoric life, are entirely conjectural and cannot be determined directly from the bones. The usual approach is to model indricotheres on the basis of living rhinos, with thick gray hairless skins with numerous folds, although we have no skin impressions or mummified specimens to test this idea, one way or another. The behavior and colors and sounds of the animals in these CG animations (such as in the documentary “Walking with Prehistoric Beasts”) are completely imaginary, and have no basis in any real-world data. Although most scientists are aware of this, a surprising number of people who watch these TV shows are stunned when they find out that so much of the show is pure fiction for entertainment, rather than science. The only real science in these shows is the interviews of expert paleontologists, and the pictures of bones and fossil localities.

Although most of the stuff you see in CG animations of prehistoric beasts (including the indricotheres) is mostly guesswork, there are living analogues that can give us some guidance about indricothere biology. The best models might be elephants, which approach indricotheres in body size. There are certain constraints about life at such large body size for elephants that must also apply to the indricotheres:

Thermoregulation: Elephants have a huge body volume and mass compared to their surface area (remember, volume increases as a cube while area only increases as a square). As the debate about hot-blooded dinosaurs back in the 1980s revealed, such huge animals with an endothermic physiology (that is, they generate their own body heat from metabolism) have a severe problem getting rid of excess body heat, especially if they live in warm climates. Living elephants have huge ears as radiators to shed excess body heat from their bloodstream, and it is reasonable to infer that indricotheres did too. African elephants and rhinos and hippos spend much of their daytime resting in the shade or wallowing in waterholes and mud puddles to cool down, and so must have the indricotheres. Elephants, rhinos and hippos feed and move mainly at night, as indricotheres must have done. Elephants and rhinos both have largely naked skin since hair holds in body heat, which is why such elephant-like naked gray skin seems appropriate for indricotheres. Large-bodied endothermic mammals are in a constant battle to dump body heat and avoid overheating.

Digestion: There are certain other constraints for large-bodied herbivores as well. All herbivores eat large amounts of cellulose in their diets, which is a relatively indigestible carbohydrate. Most plant eaters must use some kind of specialized gut bacterium in their digestive tract to break down the cellulose and release the nutrients. Such a breakdown requires fermentation, and takes time to absorb the nutrients from the fermentation process into the lining of the intestines. There are two basic types of herbivore digestion: foregut fermenters and hindgut fermenters.

The only living foregut fermenters are the ruminant artiodactyls (camels, cattle, sheep, goats, antelopes, deer, and pronghorns), which do this by “ruminating” using a four-chambered stomach. The first chamber, the rumen, is a digestive vat full of bacteria, so that when they swallow a bite of partially chewed plant material, it goes immediately into the rumen where it begins bacterial breakdown. Later, when they are resting, ruminants regurgitate some of the contents of their rumen back into their mouths, where they can “chew their cud” and break the material down further, before swallowing it again. By the time the food reaches the lining of their intestines, it is highly broken down into nutrients and easily absorbed. Thus, ruminants use nearly every bit of their food efficiently, and can survive on relatively small amounts of good-quality vegetation. But if they eat too much high-quality vegetation, they can become bloated and their rumen can swell and even rupture and kill them with all the gas released from the rapid bacterial fermentation.

The remaining herbivorous mammals are hindgut fermenters. These include the perissodactyls (odd-toed hoofed mammals, today including the horses, tapirs and rhinos), the elephants, the non-ruminant artiodactyls (pigs, peccaries, and hippos), and other herbivores such as rabbits and some primates. Instead of a highly specialized foregut with a rumen, they have the normal mammalian digestive tract, with an esophagus, acid-filled stomach, and then intestines for absorption. Most have a pouch off the intestine called a caecum that is the primary location of bacterial fermentation.  Lacking a rumen, the hindgut fermenters pass the mostly undigested cellulose through the digestive tract until it reaches the caecum, but bacterial fermentation only just starts in the caecum before the food goes through the remaining intestines and is then excreted. Consequently, they get relatively little nutrition out of each bite of fodder, and must eat much larger volumes of mostly low-quality food (especially grasses) to get enough to live on. Most hindgut fermenters, like horses, rhinos, and elephants, are by necessity be high-volume low-efficiency eaters, and eat huge volumes of material just to survive, since they are so poor at extracting the nutrients. When you see the feces of these animals (like the “road apples” of horses), they are typically full of undigested plant matter compared to the “cowpies” of a ruminant, or the tiny pellets of a deer or pronghorn. Rabbits are a special case. If you have ever kept rabbits in a hutch, you will notice that they eat their own feces. This gives them a chance to run the food through their gut a second time after the bacterial fermentation has had time to work, and get more nutrition this way.

For these reasons, there are certain things we can say with confidence about indricothere feeding dynamics. Because they were not ruminant artiodactyls, they had to be hindgut fermenters, like horses, other rhinos, and elephants, so they must have consumed and processed huge amounts of food in a day, just as elephants do now. Their feces would have been full of undigested plant material, just like those of a horse or an elephant. Like almost all large herbivores, they must have had a big part of their abdomen occupied by their large digestive tract, giving them a large bulging “gut” like that of an elephant. The fermentation in their gut, by the way, creates additional body heat, which exacerbates the problems they have of producing excess body heat to begin with.

Locomotion and Home Range: As an animal increases in body size, the stresses on their limb bones increases even more because of the power of three expansion of volume and the corresponding mass increase. Models of the dynamics of large dinosaurs show that they could not have run very fast, or their limbs would break. Modern elephants also cannot run very fast compared to true specialized runners like antelopes, horses or cheetahs. Their maximum speed in an all-out charge clocked at only 18 mph (29 kph), but their normal walking speed is about 6-12 mph (10-19 kph). Remember, they have an advantage in their speed because they have much longer limbs and strides than any other animal. Given that indricotheres were just slightly larger than modern elephants, we can predict that they too would have not been fast runners, but ambled along at a moderate pace like that of an elephant.

However, African elephants are capable to moving enormous distances (typically 20 miles or 32 km) in the course of a day, migrating from one food source to another. To support their food needs of about 300 pounds (140 kg) of food they consume in the 16 hours of each day they eat, elephants need huge home ranges of 300-600 square miles (750-1500 square km). Consequently, huge home ranges and long migrations would be expected of indricotheres as well, especially if they lived in a harsh desert scrub setting with scarce food sources that were easily wiped out. A similar model has been proposed for the large sauropod dinosaurs, which lived in a scrubby, semi-arid habitat in the Late Jurassic time (Morrison Formation), and probably roamed in small herds from one patch of trees to another.

Predators and Life Habits: Certain other ecological parameters are also dictated by the giant body sizes of elephants and indricotheres. Once they reach a large enough body size, healthy modern elephants have no natural predators—not even lions or tigers are foolish enough to tackle them. (This has all changed now with human poaching, which has nearly wiped out elephants in the wild). Only the babies and young calves are vulnerable to predators, and in elephant herds, there is a strong matriarchal hierarchy so that every calf is closely protected not only by its mother, but also by its sisters, grandmother, aunts, great-aunts, and other close female relatives. All non-human predation of elephants in the wild occurs when predators catch vulnerable calves.  Almost a quarter of the calves born to Asian elephants are lost to tigers before they reach their first birthday. If indricotheres maintained small herds in the elephant mode, such freedom from predation except for the young would also be true. However, in the Bugti beds there are gigantic crocodiles (Crocodylus bugtiensis) that are 10-11 m (33-36 feet) long! These would have been large enough to attack almost any indricothere that might be at the edge of the river to drink. Indeed, many of the specimens from the Bugti beds have crocodile tooth marks on them.

There is also a well-known relationship between the gestation period, size of the litter, and body size. Elephants have the longest gestation period of any land creature (22-24 months, or about two years). The females do not reach sexual maturity until they are ten years old, and may produce a single calf every three to four years, the slowest reproductive rate of any mammal. Such could be expected of indricotheres as well, since growing to such large body sizes, and having such large calves, is very similar to the constraints on elephant reproduction. Like elephants, indricotheres would be expected to grow quickly at first, then grow relatively slowly once they reached maturity.

There is also a strong relationship between body size, metabolic rate, and blood pressure. An elephant has a relatively slow metabolic rate. Its heart beats only 30 times per minute, while humans have a pulse of 60 beats per minute, and hamsters have a pulse rate of over 450 beats per minute! The indricothere heart would have had a pulse rate close to that of an elephant, but probably a bit higher. This is because it must have also been able to exert a blood pressure close to the 300 mm Hg that giraffes produce (humans typically have a blood pressure of 120) to be able to lift its head so far above the ground without fainting.

Finally, there is also a well-known scaling of longevity with body size, with larger animals (and their slower heart rates) living longer. A rodent typically lives no longer than 3-5 years, a cat or a goat about 15 years, pig or monkey about 20-25 years, and a cow or giraffe about 25-30 years. Elephants typically live 35-50 years in the wild (at least they did until recent years, when poaching has nearly wiped them out), and the record is 71 years. Similar lifespans could be expected of indricotheres as well.