David J Archibald. Encyclopedia of Dinosaurs. Editor: Philip J Currie & Kevin Padian. Academic Press, 1997.
The phenomenon that we call extinction is arguably one of the most misunderstood biological events. Three common misperceptions contribute to this misunderstanding. First, extinction is thought to be rare, or at least may have been rare before the explosion of the human population. Second, extinction is viewed as a negative process, one that only brings destruction. Third, the simple disappearance of a species is thought to be the same as extinction.
Estimates of both the number of living (or extant) species and the rate at which they are being driven to extinction by humankind vary widely. Estimates for the numbers of species thought to grace the earth today range from a known number of 1.4 million to tens of millions (or more). Rates of present-day extinction are placed from as low as one per year to as high as one per day, with some even higher estimates. Within the constraints of the human time-scale, extinction seems to occur at very low levels, even using the highest rates. So what if we lose one species per day if we have millions of species? If such levels continued, however, for only the extent of our Gregorian calendar—which is fast closing in on 2000 years—three-quarters of a million species would have disappeared. This would represent a loss of more than one-half of all known living species and would be very high even if the true number of extant species was closer to 10 million. How high a rate of extinction would this be compared to those in the geological past? It would rank within the top five, if not the top three, of the most significant episodes of extinction in the past 550 million years.
The process of extinction, however, was occurring well before the emergence of Homo sapiens, although we certainly have accelerated it to breakneck rates. Extinction has been ubiquitous during earth history since the origin of life approximately 3.5 billion years ago. All current estimates place the total percentage of species extinction throughout earth history at more than 90%, if not over 99%. This is a surprising figure for the uninitiated, but the surprise soon dissipates when we realize that evolution has constantly added new species throughout the past 3.5 billion years. Although there have been long intervals of increasing species diversity, life remained at a steady state for very long stretches of geologic time. The only way a steady state of species numbers could be maintained, even while evolution is occurring, is for other species to become extinct. Thus, extinction, rather than being a rare and negative event in human time frames and sensibilities, is actually a very common and positive counterpoint to evolution. Without extinction, the vast majority of extant species, including Homo sapiens, could not and would not have arisen.
I have addressed the first two common misperceptions concerning extinction; first, that extinction is thought to be rare, and second, that extinction is viewed as a negative process. However, what of the third—that the simple disappearance of a species is thought to be the same as extinction? The disappearance of species, usually based on disappearance from the fossil record, often occurs because of vagaries of the fossil record. It must also be remembered that many species are only poorly represented in the fossil record, if at all, because they lack preservable hard parts such as a shell or skeleton. If they live in regions that usually do not promote fossilization, such as in the mountains or desert, lack of a fossil record is not by itself a good measure of whether extinction has occurred. Even if we have a good fossil record for a particular species, and it disappears locally, it may continue to survive elsewhere on the globe for a considerable length of time. For example, plants such as dawn redwoods, fish such as coelacanths, and small mammals such as rat opossums (small South American marsupials) were known as fossils before living representatives were discovered growing, swimming, and scampering about elsewhere.
Finally and probably most important for the disappearance of some species is the process of speciation. During speciation a parent species may give rise to a recognizably new species while the parent species remains relatively unchanged. In other instances the parent species disappears as it splits to form two or possibly even more new daughter species. This is pseudoextinction. Unlike true extinction, which occurs when all the individuals possessing a very similar but variable genome disappear, pseudoextinction occurs when the genomes of individuals in the two daughter species are sufficiently altered so that both are clearly set on new and different evolutionary paths. This most obviously has occurred when the resulting daughter species can no longer interbreed.
If we can establish with reasonable certainty that the true extinction of one or more species has occurred, the next step is to place the extinction(s) in the context of other extinctions. Were the extinctions in question normal (or background) or were they part of a mass extinction? The major difference between normal and mass extinctions is one of scale. Because extinctions have been sampled repeatedly throughout the better known past 550 million years of earth’s history, it appears that the rate of extinction is roughly similar. Although it is difficult to provide a specific figure because of vagaries of the fossil record and differences between environments and different kinds of organisms, it can be safely said that normal extinction is below (often well below) 50%. There were intervals, however—five to be precise—during the past 550 million years that witnessed percentages of extinction well over 50%, in one case possibly reaching as high as 95%. This horrendously high level of extinction occurred at the end of the Permian Period approximately 250 million years ago. All five known mass extinctions, from oldest to youngest, were in (or at the end of) the Late Ordovician, Late Devonian, Late Permian, Late Triassic, and Late Cretaceous. There is an unresolved debate as to whether these five mass extinctions represent a separate class of extinction from normal extinctions or instead form a continuum with normal extinctions in both rate and cause.
The most recent mass extinction event near the end of the Late Cretaceous includes that of the dinosaurs, or more correctly the nonavian dinosaurs. The level of extinction for all species during the Cretaceous-Tertiary (or K-T) transition has been frequently given at approximately 75%, although there are no studies documenting this level of extinctions for species. For backboned animals or vertebrates only, the level of extinction hovers around 50% for species, but this is based on only one area, the Western Interior of North America.
Dinosaurs have come to represent not only the K-T mass extinction event for both scientists and the public but also extinction in general. Except for perhaps the hapless Dodo, nothing seems to epitomize extinction as much as dinosaurs. When we are dealing with large, relatively rare creatures such as dinosaurs, the problems of unraveling when, where, how much (magnitude), and how fast (rate) extinction occurred can be disheartening. Nevertheless, only after these questions have been addressed, even if not completely answered, can we turn to the question of possible causes.
The “common knowledge” is that dinosaurs became extinct at the same time everywhere on earth. If one is to believe not only the popular press but also some scientists, this common knowledge comes from a global record of dinosaur extinction at the K-T boundary. The idea that the dinosaurs disappeared from earth at the same time in the wink of an eye is not new; but the proposition in 1980 that an asteroid impact caused this very rapid decline and extinction of dinosaurs has given it new life. Some proponents of this theory have explicitly stated that these extinctions were essentially instantaneous around the world. Such explicit assertions about global records of dinosaurs are patently false. Most people are surprised to learn that the geographic coverage of dinosaur extinction is appallingly bad. The only region where we currently have a reasonably large sample of dinosaurs and contemporary vertebrates extending to near or at the K-T boundary (and have a fossil record of similar quality above this boundary) is in the Western Interior of North America, especially well known in the eastern part of Montana and into southern Canada. This region formed the eastern coast of a great inland sea that split North America in half in the Late Cretaceous. This was certainly an extensive region stretching for thousands of miles; nevertheless, it is a very limited region to use to explore questions of extinction on a global scale.
One study showed that of 26 dinosaur localities from near the end of the Cretaceous, 20 are from the Western Interior of North America. This represents more than 75% (20/26) of all the information we have about dinosaurs leading up to their extinction. The record of the last dinosaurs is biased not only in where the localities are found but also in the taxa and numbers of specimens. Of the 20 genera of dinosaurs known from the latest Cretaceous, 14 (70%) are known only from North America. Turning to numbers of specimens (those represented by articulated individuals composed of several bones up to a complete skeleton found in very close association), 95% of the 100 specimens known from the latest Cretaceous are from North America. These figures once more emphasize that the record of dinosaurs near the K-T boundary is almost exclusively North American.
In the next few years we may see a better global record of latest Cretaceous dinosaurs emerging. Especially promising are new finds in several sedimentary basins in China and localities in central South America. Until such time that we do have a more global record, arguments about the pace of dinosaur extinction on a global scale remain unsubstantiated speculation. For now, it must be emphasized that we simply have no record of dinosaurs that permits us to clearly show whether dinosaur extinction was catastrophically fast or glacially slow. Rather, the data we do have are more regional in scope and only permit us to examine questions of the magnitude and selectivity of these extinctions, but nothing of its pace.
Although the vertebrate record of the K-T boundary is almost exclusively limited to the Western Interior, the uppermost Cretaceous hell creek formation in eastern Montana has yielded a taxonomically rich sample of 107 vertebrate species. The record includes 12 major vertebrate lineages—5 species of sharks and relatives, 15 of bony fishes, 8 of frogs and salamanders, 10 of multituberculate mammals, 6 of placental mammals, 11 of marsupial mammals, 17 of turtles, 10 of lizards, 1 of the unfamiliar champsosaurs, 5 of crocodilians, 10 of ornithischian dinosaurs, and 9 of saurischian dinosaurs (not including birds). Of these 107 vertebrate species or species, 49% (52 of 107) survived across the K-T boundary in the Western Interior. This is the minimum percentage survival of vertebrate species across the K-T boundary because some of the very rare species may have survived undetected. Twenty of the 107 are quite rare species, represented by fewer than 50 identifiable specimens out of 150,000 specimens estimated to have been recovered from the Hell Creek Formation. Although an accurate estimate is not possible, certainly some of these very rare species must have survived. The extreme and very improbable case would be if all 20 survived. This would be 67% (72 of 107) survival. This provides the extreme maximum percentage survival in the region. An educated guess, though no more than a guess, would be that no more than 60% of vertebrate species survived the K-T boundary.
When examined in greater detail, a very interesting pattern emerges within the record of survival and extinction for these Hell Creek Formation vertebrates. This is a pattern of disparity in species survival among the 12 major groups. Only 5 of the groups— sharks and relatives, marsupials, lizards, ornithischians, and saurischians—contributed to 75% of the extinction. What do sharks, lizards, marsupials, and the two lineages of dinosaurs have in common in these faunas, other than that each suffered at least 70% or more species extinction at the K-T boundary in western North America? If we are to understand causes of extinction at the K-T boundary we must explain this disparate pattern of extinctions. Any theories about the cause(s) of extinction of dinosaurs and their contemporary vertebrates must be able to explain the previously discussed pattern. Any theories must explain why sharks, lizards, marsupials, ornithischians, and saurischians suffered very high levels of or even total extinction, whereas bony fishes, frogs and salamanders, multituberculate mammals, placental mammals, turtles, champsosaurs, and crocodilians suffered 50% or often much less extinction.
There are as many as 80 theories or variants of theories that have been proposed for dinosaur extinction. Many of these are either frivolous (Martians hunted the dinosaurs to extinction) or untestable (a plague spread through dinosaur populations). Three particularly important and testable hypotheses are the impact, volcanism, and marine regression theories. To these recently has been added another called the Pele theory that suggests that, among other things, the decrease in atmospheric oxygen would have had a detrimental affect on dinosaurs. We must await further information to test this last theory adequately. Only the impact and marine regression theories have been relatively thoroughly tested with the vertebrate fossil record by their respective proponents, although proponents of the volcanism theory suggest that many of the biotic responses to an impact would also be found with massive volcanism. One method to assess the efficacy of these theories is to examine them in the context of the known K-T vertebrate record, starting with the impact theory.
The original scientific paper in 1980 advocating the impact theory still offers the basic mechanism of how such an impact might cause extinction among both animals and plants, including vertebrates. The impact would create a dust cloud enveloping the globe for a few months to a year. Darkness would shroud the world as long as the dust remained in the atmosphere. Photosynthesis in the sea and on land ceased. As the plants died or became dormant, herbivores soon starved, followed by the carnivores.
Some of the best physical evidence of such an impact is the discovery of anomalously high levels of the rare earth element iridium at the K-T boundary and the probable remains of an impact crater. Although iridium is found on the earth, especially deep in the earth, high levels of iridium are associated with extraterrestrial events. Generally, the larger the impacting object, such as an asteroid, the greater the increase in signature elements such as iridium. The remains of a probable impact structure dubbed “Chicxulub,” approximately 110 miles across at the northern tip of the Yucatan peninsula from at or near the K-T boundary, are strong evidence for an impact. Neither elevated levels of iridium nor an impact crater, however, are direct evidence for specific causes of extinction at the K-T boundary.
An incorrect assumption often made in testing the impact theory and its possible corollaries is that all major taxa show very high levels of extinction across the ecological spectrum on a global scale. As already discussed, for many organisms, but most notably dinosaurs and their contemporary vertebrates, there is no such global record at the species level. The impact- generated scenario of extremely high levels of catastrophic extinction across most environments is so broad spectrum and tries to explain so much that it is difficult to test. The burden of proof for sweeping, catastrophic extinction scenarios rests with the proposers of the theory. The various corollaries of the impact theory, such as a sudden cold snap, highly acidic rain, or global wildfires, are more easily tested using the known K-T vertebrate record.
A short, sharp decrease in temperature was not emphasized in the originally proposed hypothesis, but it soon became an important corollary of the impact theory. It is argued that if tremendous amounts of dust were injected into the atmosphere after a large impact, the darkness would not only suppress photosynthesis but also produce extremely cold temperatures. This hypothesized condition has become known as “impact winter.” It is argued that following a large impact, ocean temperatures would decrease only a few degrees because of the huge heat capacity of the oceans, but on the continents, however, temperatures would be subfreezing from 45 day up to 6 months. The temperature would remain subfreezing for about twice the time of darkness caused by the dust.
If a suddenly induced, prolonged interval of subfreezing occurred in subtropical and tropical regions today, which vertebrates in this climate, which is similar to that of the latest Cretaceous, would be most affected? In general, ectothermic tetrapods would suffer most. Ectotherms, as the name suggests, heat or cool themselves using the environment. Endotherms such as mammals and birds generate their heat through metabolic activity. In endotherms approximately 80% of food consumption goes toward thermoregulation (regulation of body temperature).
Fishes, which are by and large ectothermic, would be generally less affected by a severe temperature drop.
Today, the northern limit of turtles and crocodilians is controlled by temperature. These animals cannot tolerate freezing, becoming sluggish or immobile at 10-15+C. Various amphibians and reptiles do inhabit areas with low winter temperatures or severe drought, but they have evolved methods of torpor (estivation and hibernation) to survive. These are the exceptions, however, because species diversity for ectothermic tetrapods is far higher in warmer climates. More important, we should not assume that Late Cretaceous ectothermic tetrapods living in subtropical to tropical climates such as in eastern Montana were capable of extended torpor. Torpor is most often preceded by decreases in ambient temperature, changes in light regimes, and decreases in food supply. The ectotherms in eastern Montana could not have anticipated a short, sharp decrease in temperature. This is true even if the impact had occurred during a Northern Hemisphere winter when temperatures would been slightly lower. We must remember that this was a subtropical to tropical setting, and thus the extended, subfreezing temperatures advocated by proponents of this corollary would have been devastating to ectotherms even during a terminal Cretaceous winter in Montana.
Except for a 70% decline in lizards, ectothermic tetrapods (frogs, salamanders, turtles, champsosaurs, and crocodilians) did very well across the K-T boundary. The corollary of a sudden temperature decrease simply does not fit with the vertebrate data at the K-T boundary. A latest (but not terminal) Cretaceous vertebrate fauna from northern Alaska strengthens the evidence that a hypothesized sudden temperature drop was not a likely cause of K-T boundary extinctions. Comparing the Late Cretaceous vertebrate faunas from Alaska and eastern Montana reveals a striking difference. Although the Alaskan fauna is decidedly smaller and with fewer species than that from eastern Montana, both have sharks, bony fishes, dinosaurs, and mammals. The Alaskan fauna, however, completely lacks amphibians, turtles, lizards, champsosaurs, and crocodilians. These taxa comprise 41 of 107 (38%) of the eastern Montana fauna. If even the fairly balmy temperature range of 2-8+C for Late Cretaceous Alaska was enough to exclude ectothermic tetrapods, a severe temperature drop to below subfreezing temperatures at the K-T boundary should have devastated the rich ectothermic tetrapod faunas at midlatitudes. These species flourished.
A second prominent corollary of the impact theory is highly acidic rain. The most commonly cited acids as products of an impact are nitric and sulfuric acid. It is argued that nitric acid would be produced by the combination of atmospheric nitrogen and oxygen as a result of the tremendous energy released by an impact. Sulfuric acid would be produced because large amounts of sulfur dioxide are vaporized from rock at the impact site. These acids would be precipitated in the form of rain. Estimates of the pH of these acid rains vary, but estimates reach as low as 0.0-1.5! It is suggested that global effects could have caused the pH of near-surface marine and fresh water to below 3. In today’s environment, rain below a pH of 5.0 is considered unnaturally acidic. Rain as low as 2.4 has been recorded, but annual averages in areas affected by acid rain range from 3.8 to 4.4. Acid fogs and clouds from 2.1 to 2.2 pH have been recorded in southern California and have been known to bathe spruce-fir forests in North Carolina. The biological consequences of such low pH values vary from one vertebrate group to another but are always detrimental. Aquatic species (fish, amphibians, and some reptiles) are the first and most drastically affected, with those reproducing in water being the first to suffer. If pH becomes lower than approximately 3.0, adults often die. The effects on aquatic vertebrates across the K-T boundary would have been very bad if a pH of 3.0 was reached and truly horrendous if it hit 0.0 as suggested by some authors. Some advocates of K-T acid rain argue that the surrounding soils or bedrock would have buffered the aquatic systems, and they even suggest that limestone caves could have been important refugia for birds, mammals, amphibians, and small reptiles. The only problem with this scenario is that there were none of these kinds of buffering soils or bedrock or limestone caves in eastern Montana in the Latest Cretaceous. Based on what we know of our modern biota’s reaction to acid rain, aquatic animals should have been devastated by acid rain at the K-T boundary. Of all the aquatic species, only sharks and their relatives show a drastic drop in eastern Montana. Thus, the likelihood of low pH rain is highly implausible.
A third corollary of the impact theory that receives various levels of support is global wildfire resulting from the aftermath of the impact. Soot and charcoal have been reported from several sites at the K-T boundary coincident with the enrichment of iridium noted earlier. It was argued that this pattern is unique and must come from the extremely rapid burning of vegetation equivalent to half of all the modern forests! Other scenarios argue that approximately 25% of the aboveground biomass burned at the end of the Cretaceous!
Such a global conflagration is really beyond our comprehension. In order to grasp the magnitude of this scenario, imagine one-quarter to one-half of all structures on the globe engulfed in flames within a matter of days or weeks. This still would be only a fraction of what is argued to have been burned at the K-T boundary. In such an apocalyptic global wildfire, much of the aboveground biomass all over the world would have been reduced to ashes. In fresh water, those plants and animals not boiled outright would have faced a rain of organic and inorganic matter unparalleled in human experience. These organisms would have literally choked on the debris or suffocated as oxygen was suddenly depleted with the tremendous influx of organic matter. The global wildfire scenario is so broad in its killing effects that it could not have been selective, but, as discussed earlier, the vertebrate pattern of extinction and survival is highly selective. Thus, it is no surprise that this scenario of equal opportunity losers does not show any significant agreement with the pattern of vertebrate extinction and survival at the K-T boundary.
Not only is there almost no fossil evidence supporting global wildfire but also the physical basis for such an event is suspect. It is argued that there is a global charcoal and soot layer that coincides with the K-T boundary, whose emplacement is measured in months. This also assumes that the sedimentary layer encasing the charcoal and soot was also deposited in only months. This is demonstrably not the case for at least one K-T section that continues to be cited in these studies—the Fish Clay of the Stevns Klint section on the coast of Denmark. The Fish Clay is a laterally discontinuous, complexly layered and burrowed clay reflecting the conditions at the time of its deposition. It is not the result of less than a year of deposition caused by an impact-induced global wildfire. Thus, carbon near the K-T boundary at Stevns Klint as well as in other sections is likely the result of much longer term accumulation during normal sedimentation.
When all of the corollaries of an impact of an asteroid or comet are compared to the pattern of extinction and survival for vertebrates at the K-T boundary in eastern Montana, there is relatively poor agreement. Without special pleading, these corollaries as currently proposed are unlikely causes of vertebrate extinction. This does not mean that all corollaries of an impact should be rejected, but it is imperative that those proposing the different corollaries separate those that are supported by the vertebrate fossil record from those that are not.
The next hypothesis is the volcanism theory. Although some proponents of the impact theory do not agree, many advocates of both theories feel that a number of the same physical events would have occurred at the K-T boundary if either extensive volcanism or an impact took place. They also say that the biological results would be similar. Given the previous discussion of how most of the corollaries of the impact theory do not test well against the vertebrate fossil record, the volcanism and impact theories are equally weak in their biological predictions. The major difference in these two theories is in their timing. Whereas the impact theory measures most of the cataclysmic effects in months or years, with physical effects possibly lingering for a few hundred or a few thousand years, the volcanism theory measures effects into the millions of years. The effects of many volcanic eruptions, such as that of Mt. St. Helens, linger for a few months or a few years. Many other episodes of volcanism are very prolonged. These are flood basalt eruptions. The best known in the United States is the 16-million-year-old Columbia River flows in the northwest. In the past 250 million years, arguably one of the biggest flood basalt eruptions occurred on the Indian subcontinent. This was occurring during (and is probably related to) the collision of the subcontinent with the remainder of Asia. Its most obvious manifestation today is the tallest mountain range in the world, the Himalayas. These flood basalts, known as the Deccan Traps, cover an immense part of both India and Pakistan. Individual flows in the sequence cover almost 4000 square miles with a volume exceeding 2400 cubic miles. Individual flows average 30-160 ft thick, sometimes reaching 500 ft. In western India the accumulations of lava flows is 7800 ft, or 1.5 miles thick. The flows originally may have covered almost 800,000 square miles with a volume possibly exceeding 350,000 cubic miles (an area the size of Alaska and Texas combined, save about 30,000 square miles, to a depth of more than 2000 ft). Based on radiometric dating, paleomagnetics, and vertebrate fossils, the bulk of the eruptions are centered around the K-T boundary, during a reversal in earth’s magnetic poles known as 29R or 29 reversed. The number 29 represents the 29th reversal of the earth’s magnetic field counting backwards from the present, which today has normal polarity by definition. The K-T boundary happens to fall in 29R.
What would the effect have been on the global biota if something the magnitude of the Deccan Trap erupted for tens of thousands of years or longer? One of the greatest effects would have been to increase and maintain a much higher level of particulate matter in the atmosphere. Whether it would have caused warming through a greenhouse effect, cooling because of less light, or simply prettier sunsets is not certain. The amount of CO2 pumped into the atmosphere by the eruptions may have been a boon for green plants that require CO2 for photosynthesis, but a reduction in light reaching the surface because of particulate matter may have canceled the effects of increased CO2. The effects of added particulate matter might have prevailed for no other reason than that they would linger longer after eruptions ceased, whereas the release of CO2 would have diminished much more rapidly after each eruption stopped.
If the latter scenario is correct, the longer term effects over a million years or more would be to push a global cooling. Most estimates suggest that regional if not the global climate cooled through the K-T transition. Because the time frame is moderately long, many species, especially smaller ones, on land or in the sea could adapt to changes, whereas larger species such as dinosaurs may not have been as fortunate. Although it probably was not a cause of extinction for most species, the cooling across the K-T boundary would have been an added stress.
A final long-term effect suggested for eruption of the Deccan Traps is reduced hatching success for eggs of herbivorous dinosaurs. Volcanic activity can release elements such as selenium that are highly toxic to developing embryos. Increased levels of selenium in the eggshells of dinosaurs are known from near the K-T boundary in southern France. Poisoning of eggs has also been reported from dinosaur eggs near the K-T boundary in Nanxiong Basin, southeastern China.
The final hypothesis that has been tested with the vertebrate fossil record is the marine regression/habitat fragmentation theory, or simply marine regression theory. Many areas of the terrestrial realm were repeatedly inundated by shallow epicontinental seas throughout geologic history. The term “epicontinental” refers to the occurrence of these very shallow seas upon the continental shelves and platforms rather than in deep ocean basins. Epicontinental seas reached depths of only 1500-2000 ft, very shallow compared to most large modern marine bodies. Epicontinental seas are almost nonexistent today, except for such bodies of water as Hudson Bay. It is known that during the Late Cretaceous, large areas of continents were submerged under warm, shallow epicontinental seas.
It became clear only recently just how dramatic the loss of these seas was leading up to the K-T boundary. There is absolutely no mistaking that the K-T loss of shallow seas (or increase in nonmarine area) is greater than at any time in the past 250 million years. The nonmarine area increased from 42 million m2 to 53 million m2—more than a 25% increase. This is the equivalent of adding the land area of all of Africa, the second largest continent today. The second largest increase in continental area in the past 250 million years occurred across the Triassic-Jurassic boundary. Like the K-T transition, this is also during one of the five universally recognized mass extinctions during the Phanerozoic or last 550 million years.
Some of the most dramatic additions of nonmarine areas at or near the K-T boundary occurred in North America. Near the end of the Cretaceous, maximum transgression divided North America into two continents. As regression continued until at or near the K-T boundary, coastal plains decreased in size and became fragmented; stream systems multiplied and lengthened; and as sea level fell, land connections were established or reestablished.
The driving force for these repeated inundations or transgressions of the lower-lying portions of continents is still not fully understood. The general consensus is that it is related to plate tectonics. It is thought that rises in sea level and inundations began as the motion of the plates increased. As this occurs, the margins along which the colliding plates converge are subducted or pushed downwards into the earth. This causes inundation by the seas upon shallow continental shelves and platforms.
Whatever the geophysical factors driving the process, the physical manifestations of marine regression, like impacts and volcanism, are important ultimate causes of extinction. Although these processes of marine transgression and regression were global in extent, a closer examination of North America is best because, as I have emphasized, this is where we have the vertebrate data at the K-T boundary. North America was split into two continents—a western continent (Laramidia) and an eastern continent (Appalachia)—by the Pierre Seaway for almost 40 million years during the Late Cretaceous. Most of our latest Cretaceous vertebrate fossils come from the east coast of Laramidia. The west coast of Appalachia as well as the eastern seaboard of Appalachia have also produced some specimens.
In the last few million years of the Cretaceous the Pierre Seaway began to regress from both Laramidia and Appalachia. At or just shortly before the K-T boundary, the seaway reached its nadir. Placement of the receding coastlines both north and south have not been well established, but we know the southern coastline reached well into Texas. There is no question that there was a dramatic reduction in coastal plains. This is exactly the kind of environment from which we are sampling the last of the Late Cretaceous vertebrate community. A common refrain is that because the total amount of land increased with the regression, dinosaurs should have had more, not less, area and more environments in which to live. We know with considerable certainty that dinosaurs did live in other environments such as the higher, drier Gobi Desert in Mongolia during part of the Late Cretaceous. Currently, however, the only well-known vertebrate communities that preserve dinosaurs at the K-T boundary are coastal. Thus, arguments about what dinosaurs and other vertebrates may or may not have done in other environments are moot. It is simply incorrect to say that the dinosaurs and other vertebrates may have survived elsewhere when we have little or no information about other environments.
The drastic reduction of coastal plains put tremendous pressure on some, especially large vertebrate species. Reduction of habitat, for example, in the Rift Valley system of East Africa, today first affects larger vertebrates, especially mammals. In the shrinking coastal plains of latest Cretaceous North America, the equivalent large vertebrates first affected were the dinosaurs. An additional problem, whether in East Africa today or the coastal plains of latest Cretaceous North America, is the fragmenting of the remaining habitat. This process, as a result of human activity, has become known as habitat fragmentation. In larger, undisturbed habitats, animals (and plants) can spread more freely from one area to another. If the habitat is fragmented, although the amount of habitat may not have been greatly altered, it will reduce the flow of species from one fragment to another.
For some species, even seemingly small barriers such as two-lane roads can be insurmountable. The results can be disastrous if viable populations cannot be maintained in the various fragments. Fragmentation can lead to extinctions. Barriers also arise in nature even among animals that would seem easily capable of dispersing. Although no doubt the result is very often extinction, we usually only see what survives in the form of differences between closely related species. Small arboreal primates in the rainforests of both South America and Africa form small fragmented groups that are isolated from each other often by rivers only tens of yards wide. Another example is the Kaibab squirrel on the North Rim of the Grand Canyon. Unlike its nearest relative, Abert’s squirrel, which is found on the south side of the Grand Canyon and in the western United States and Mexico, the Kaibab squirrel is restricted to an area of only 20 + 40 miles. Fragmentation, in this case the development of the Grand Canyon, helped produce the differences, but the margin between this and oblivion for the Kaibab squirrel is slim.
The idea of habitat fragmentation not only extends to natural processes operating today but also to processes operating in the geological past. Although historical habitat fragmentation is not well understood by earth scientists, it is an all too real phenomenon among biologists studying the effects of human activity in modern rainforests and in urban settings. Declines of bird and mammal populations have been well documented in the city of San Diego as urban development divides and isolates habits in canyon areas. One would not expect that the natural equivalent of habitat fragmentation would be easily, if at all, preserved in the rock. The forcing factor for habitat fragmentation in the latest Cretaceous—marine regression—is a thoroughly documented fact during the waning years of the Late Cretaceous in North America. Globally, marine regression occurred within this same general time frame although how close in time it occurred in various regions is a matter of debate.
Theory predicts that large species would be the most severely affected by habitat fragmentation for the reasons discussed previously. During the K-T transition in eastern Montana, only 8 of 30 large species survived and these are partially or entirely aquatic (2 fishes, 1 turtle, 1 champsosaur, and 4 crocodilians), whereas all 22 large terrestrial species (and 1 aquatic species) became extinct (1 turtle, 1 lizard, 1 crocodilian, and 19 dinosaurs). Thus, predictions from habitat fragmentation fit the observed data very well.
As noted previously, two other major physical events occur with marine regression in addition to habitat fragmentation—stream systems multiply and lengthen and, as sea level falls, land connections are established or reestablished.
Only a few of the K-T boundary stream systems have been studied in detail in the Western Interior, and thus we do not know the exact drainage patterns for most latest Cretaceous and early Tertiary stream systems in the eastern part of Laramidia. Nevertheless, we are certain that as new land was added following marine regression in the early Tertiary, stream systems increased and lengthened. This process is another major corollary of marine regression. When freshwater habitats were bolstered following marine regression, most aquatic vertebrates did well, except those with close marine ties—sharks and some bony fishes. Such fishes may need to spend at least a portion of their life in a marine environment, in some instances to reproduce. The major group most likely to suffer would have been the sharks and their relatives. In fact, all five species of sharks disappeared. It is not clear, however, whether these disappearances from the Western Interior are actually extinctions at the K-T boundary or whether the shark species survived elsewhere in marine environments into the earliest Paleocene. The problem is that the definitively oldest marine sediments that postdate the K-T boundary in the Western Interior are no older than late early Paleocene in age. This means that there is a gap in marine sedimentation in the Western Interior of possibly 1 million years or more immediately after the K-T boundary. This pattern of disappearance and reappearance strongly suggests that as the Pierre Seaway regressed further and further away from eastern Montana, all sharks and relatives departed because connections to the sea became attenuated. New species of elasmobranchs did not occur in the area until a smaller transgression reached the Western Interior at or just before middle Paleocene times. This is known as the Cannonball Sea, which was a smaller seaway than the Pierre. The total disappearance of sharks and relatives is the only prediction that can be made with any certainty as a result of the loss of marine connections and the lengthening of stream systems. The increase of stream systems was a positive factor helping to mitigate other stresses that may have been put on the freshwater system.
New land areas were exposed as sea level lowered. In some cases this included the establishment or reestablishment of intercontinental connections. One such connection was the Bering land bridge joining western North America and eastern Asia. At various times during the Late Cretaceous this bridge appeared and then disappeared. This is suggested by similarities in parts of the Late Cretaceous vertebrate faunas in Asia and North America, especially the better studied turtles, dinosaurs, and mammals. Competition and extinction often result from biotic mixing, but predicting the fates of various taxonomic groups is usually not possible. An exception may have been the fate of marsupials in North America near the K-T boundary.
The oldest marsupials are known from approximately 100-million-year-old sites in western North America. By some 85 Mya, we know of about 10 species of marsupial. This rose and stayed at about 15 species from approximately 75 Mya until the K-T boundary approximately 65 Mya, when it plummeted to one species. These were all quite small mammals, from the size of a mouse up to a very well-fed opossum or raccoon. Their teeth were very much like those of extant opossums, with slicing crests and well-developed but relatively low cusps (compared to contemporary placental mammals) for poking holes in insect carapaces, seeds, or whatever they found. Most did not appear to be specialists on any particular food. With the reestablishment of the Bering land bridge (or at least closer islands) near the K-T boundary, a new wave of placental mammals appeared in western North America. These mammals (traditionally known as condylarths) were the very early relatives of modern ungulates and whales. Their appearance in North America coincides with the very rapid decline of marsupials near the K-T boundary. Within a million years of the K-T boundary, 30 species of these archaic ungulates are known in North America, and their numbers kept on rising. Our best guess now is that the lineage that gave rise to these mammals first appeared in middle Asia between approximately 80 and 85 million years ago and reached North America near the K-T boundary. What is of interest is that the archaic ungulate invaders had dentitions very similar to contemporary marsupials and presumably ate similar things. Its seems more than coincidence that marsupials did well in North America for approximately 20 million years only to almost disappear with the appearance of the ungulate clade. It is ironic that both marsupials and ungulates were joint invaders of South America very soon after the K-T boundary. Their dentitions were already beginning to show differentiation, with the marsupials headed toward carnivory and ungulates headed toward herbivory. It shows what a little cooperation can do.
These various physical events accompanying marine regression fit very well the pattern of extinction and survival described previously for the 107 vertebrate species from near the K-T boundary in eastern Montana. In fact, the patterns of extinction and survival for 11 of 12 of the major vertebrate groups agree very well with predictions from the marine regression theory discussed previously.
Marine regression, extraterrestrial impact, and massive volcanism are all major environmental events that occurred near or at K-T boundary. None of these physical events appears sufficient by itself to be crowned the sole cause of extinctions at the K-T boundary. The evidence outlined here overwhelmingly supports this view. What is less certain is whether all three were necessary for the pattern of extinctions we see at the K-T boundary. No one knows for sure. Both marine regression and an impact were apparently necessary to give us the pattern of turnover at the K-T boundary for all species, not just the vertebrates discussed in this essay. The role of volcanism or the effects of the Pele hypothesis are less certain. The effects of massive volcanism are formidable, but the purported biological effects have not been as closely explored as those of the other two events. The Pele hypothesis simply has not yet been properly studied and tested.
What is very clear is that at least three physical events—marine regression, extraterrestrial impact, and massive volcanism—did coincide near the K-T boundary, but this does not mean that all three events necessarily occurred simultaneously at this boundary. The greatest addition of nonmarine area in the past 250 million years brackets the Late Cretaceous mass extinctions. One of the largest known impact craters is thought to have been identified at very near the K-T boundary. Massive volcanism had been pouring out great quantities of lava sporadically for several million years during the K-T interval. This was clearly one of the most geologically complicated and biologically challenging episodes in earth history. Given these challenges it is surprising that more species did not succumb. It suggests that life has been more resilient than we dreamed, or have a right to hope for, given the stresses we are placing it under today.