Hen’s Teeth and Horse’s Toes

by Stephen Jay Gould

  The major theories of mass extinction can be divided into two groups according to their stance on each of two issues: source (within or outside the earth) and rate (truly sudden or only relatively rapid). The earth itself is a source for some proposed causes, either by such physical mechanisms as changing climates engendered by shifting continents, or such biological factors as disease, competition, and collapse of food chains. Extraterrestrial hypotheses have ranged from variation in solar output, to cosmic radiation from nearby supernovae, to impacts of various bodies. Speaking of rate, some theories posit not merely a relatively rapid blip in the vastness of time but true cataclysms, disasters on the short scale of a human life—impacts of extraterrestrial bodies, for example. Other theories invoke processes that would be too slow to note during a human lifetime, but do their work in thousands or even millions of years, against a backdrop of billions. Most of these noncatastrophic theories implicate changes of climate, including drops in sea level and the growth of glaciers.

  Iridium, Geologists, like all folks, have their prejudices. They prefer causes emanating from their own domain, the earth. Since Lyell’s day they have been trained to view major change as the accumulation of small inputs based on processes that can be observed in the relatively calm geologic present. These preferences have combined to give cataclysmic extraterrestrial theories a poor shake. Yet I think that few geologists would deem it inherently impossible, or even unlikely, that the earth might have suffered grievous cosmic insults at infrequent intervals during its vast history.

  But another reason, better than traditional prejudice, governs the low esteem of extraterrestrial catastrophes. Geologists have not known, even in principle, any way to obtain direct evidence for them. What direct sign would a supernova or pronounced variation in solar intensity impose upon the earth? Indeed, the traditional argument for zapping by cosmic rays from supernovae relies upon total lack of evidence—the fact of mass extinction accompanied by no recognized geologic agent that might have caused it! Thus, many geologists, including myself, have long found themselves in the uncomfortable position of viewing extraterrestrial catastrophes as inherently plausible but rooting strongly against them. For what good is a theory, even a correct theory, that can generate no confirming evidence? The asteroidal theory has changed all that.

  The facts of the Cretaceous extinction exert constraints upon the types of theories we may propose to explain them. We know, for example, that the extinctions occurred throughout the world and in all major environments—land, air, and sea. This fact alone virtually invalidates the entire panoply of popular theories that would attribute the extinction of dinosaurs to a cause related only to their supposed lumbering inefficiency—mammals eating their eggs, flowering plants pumping too much oxygen into the atmosphere, hyperpituitarism arising from large size and leading to sterility. Any harebrained idea can win notoriety in a context of such public fascination. Someone once proposed in all seriousness that male dinosaurs simply became too heavy to mount their partners for sexual intercourse, although I could never figure why little Velociraptor became extinct along with its giant cousins (not to mention what the giant brontosaurs were doing during the 100 million years or so of their success). The primary fact of dinosaur extinction is its timing as part of a global mass dying. We need a general theory, not a set of facile speculations for single groups.

  We also know that the Cretaceous extinction included some aspects of geologically sudden death and others of more lingering demise. For some groups, the final phase of the Cretaceous seems to have acted more as a coup de grâce than an exterminating angel. Dinosaurs and ammonites had been in decline for millions of years. The dinosaur fauna of the latest Cretaceous did not include one of everything around a waterhole (as the multicolor chart on my kid’s wall suggests), but a sharply reduced assemblage consisting largely of Tyrannosaurus, Triceratops, and a few smaller creatures. We can also correlate these slower declines with some geological events often implicated in extinctions (but please remember that correlation need not imply cause). Sea level declined steadily throughout the late Cretaceous; the continuous seaway that had split North America in two, running from Alaska to the Gulf of Mexico, withdrew gradually in both directions. As sea level dropped and continents grew in height and extent, temperatures began a general decline that continued throughout the next seventy million years, culminating in our recent (and still uncompleted) cycle of glacial ages.

  Falling sea level has accompanied nearly every mass extinction that the earth has suffered; this correlation is about the only aspect of mass extinction that evokes general agreement among geologists. Its negative effect upon biological diversity also makes sense—for falling seas drained the extensive but shallow continental shelves, thereby removing a large chunk of living space from the domain of shallow-water invertebrates, the dominant fauna of our fossil record. Harsher conditions then spread across the land as the increasingly erratic and generally colder weather of a more “continental” earth prevailed. I doubt that any dinosaur ever ate an ammonite (although the giant mosasaurs, overgrown varanid lizards, did), but the coordinated decline of both groups may be causally related to dropping sea levels.

  Yet we cannot attribute the entire Cretaceous extinction to a gradually deteriorating climate. Something more dramatic must have happened, as the plankton line testifies. Perhaps this dramatic cause gained greatly in effect because more groups than usual were in decline and therefore susceptible to a coup de grâce. In this sense, any complete account of the Cretaceous extinction will probably include a complex combination of dramatic end superimposed upon general deterioration.

  In any case, geologic evidence constrains us to look for a contributing cause that is worldwide in effect, able to exterminate groups in all major habitats, and geologically sudden for at least some of its results. Which brings me back to asteroids.

  The asteroidal theory, like so many interesting hypotheses in science, had its root in a study with markedly different aims (you cannot actively look for the utterly unexpected). A team headed by Luis and Walter Alvarez at Berkeley, California, thought that they might use the amount of iridium in sediments as an indicator of their depositional rate. Iridium, a rare metallic element of the platinum group, is one to ten thousand times more abundant in asteroids and meteorites than in the earth’s crust and upper mantle. (Since both the earth and meteorites congealed from the same source, we must assume that the earth as a whole contains as high a percentage of iridium as the meteorites. But the earth melted and differentiated, and such heavy, unreactive elements as iridium sank into the inaccessible central core. The smaller bodies that form meteorites and asteroids never differentiated and therefore maintain iridium in its primeval abundance.) Hence, most iridium in the earth’s sediments comes from extraterrestrial sources. Working with the common assumption that meteorites and cosmic dust fall upon the earth in a fairly constant rain, the Alvarezes reasoned that sediments high in iridium must have formed slowly since relatively less earth-based debris had accumulated to dilute the cosmic influx.

  But the Alvarezes were not prepared for the anomalously high concentrations of iridium that they found in two places: in the Umbrian Apennines of Italy and near Copenhagen. Iridium levels were 30 times higher than average in Italy and 160 times higher in Denmark. Moreover, an analysis of twenty-seven other elements in the Italian sample showed that none departed by more than a factor of 2 from “average behavior” in ordinary sediments. The anomaly involves iridium alone.

  The Alvarezes wondered if they could apply the style of explanation they had originally favored: could sedimentation have been slow enough in these two places to yield such a high concentration of iridium from the normal cosmic rain alone? But they could find no evidence or even think of any reason for believing that these sediments had formed during a virtual shut-off of normal, depositional processes in the ocean. Instead, they were forced to reverse their perspective: sedimentation had been more or less normal; the iridium repres
ented a genuine cosmic influx of unusual amount, not an undiluted gentle rain. The Alvarezes had another outstandingly good reason to favor such a reversal. Both samples came from thin clays deposited at the very top of the Cretaceous—coincident with the great extinction.

  But what extraterrestrial source might have both produced the iridium and acted as a cause of the great extinction? The Alvarezes looked first to that venerable old standby of cosmic theories—the supernova that exploded near the earth and zapped our planet with so much cosmic radiation that many creatures mutated themselves out of existence. Yet, after flirting with the idea (partly in public) for a while, they dropped it—to my great delight, for it has never made any biological sense to me, despite its almost knee-jerk popularity in the “disaster literature.”

  Radiation increases the mutation rate and yields a population with more variation. But more variation per se leads neither to extinction by prevalence of monstrosities nor to unusually rapid rates of evolution, because evolutionary tempos seem to be controlled by a different force—natural selection. Ordinary populations possess enough variation (without external goosing) to permit evolutionary rates so rapid that they appear instantaneous in geologic perspective. Mutation rates so high that they kill animals directly (not through the passage of defective genes to offspring) require supernovae too close to our sun to be plausible, given the spacing of stars in our part of the galaxy.

  The Alvarezes now cite three reasons for rejecting a supernova:

  A supernova would also have produced a high concentration of an ion of plutonium (244Pu), yet the Italian and Danish clays contain levels of this ion ten times below the predicted value for a supernova.

  Iridium occurs as two common isotopes (191Ir and 193Ir). Since all the objects of our solar system had a common origin, the ratio of these two isotopes should be the same in meteorites and in the small amount of iridium indigenous to the earth’s crust. Iridium formed in other stars might exhibit a different ratio. The ratio in the anomalous Italian and Danish samples matches that of the earth’s indigenous iridium and probably came from our solar system.

  In order to zap the earth with as much iridium as the Italian and Danish samples contain, the exploding star would have to be so close to our sun that the probability of such an event becomes too small to be believed.

  Since the ratio of iridium ions led the Alvarezes to seek a source within our own solar system, they turned to objects that might hit the earth with reasonable probability and do sufficient damage. Most asteroids orbit the sun in the large space between Mars and Jupiter, but a few follow more erratic paths, and some, called Apollo objects, cross the earth’s orbit in their wanderings. Since the asteroid Apollo was discovered in 1933, twenty-seven others that cross the earth’s orbit have been sighted. Astronomers discover an average of four more each year, while two independent estimates yield about 700 for the probable number of Apollo asteroids more than one kilometer in diameter. The Alvarezes conclude that occasional collisions between Apollo asteroids and the earth are inevitable.

  In short, their scenario for the Cretaceous extinction calls upon the impact of an Apollo asteroid approximately ten kilometers in diameter. They calculate that such an object would have produced a crater more than 150 kilometers in diameter and injected so much dust into the atmosphere both from its own pulverization and by dismemberment of the earth around it that our entire planet became as dark as Egypt during Pharaoh’s ordeal. Photosynthesis might have been completely suppressed for a decade or so, leading to immediate death of the photosynthetic plankton (with their short generation time measured in weeks) and a subsequent collapse of the oceanic food chain based upon them. Most species of large terrestrial plants might have survived through the dormancy of their seeds, but the parental plants themselves would have died and taken their dinosaur herbivores with them. The Alvarezes calculate that Apollo asteroids of this size could have hit the earth with sufficient frequency to cause the five major extinctions that have punctuated the history of life since the inception of an adequate fossil record some 600 million years ago.

  This scenario also contains major problems. I am most disturbed by the special pleading required to make the pattern of extinctions come out right. I can buy the oceanic argument, but balk at the Alvarezes’ attempt to explain why three terrestrial groups got through the great darkness relatively unscathed: plants, small vertebrates (mammals and birds), and nearshore vertebrates. In a classic case of eating and having their cake, they permit mammals to survive by eating seeds, yet call upon those very seeds to save the plant species that engendered them. Their explanation for nearshore vertebrates is grounded more in hope than in logic: survival “may be due to their ability to utilize food chains based on nutrients derived from decaying land plants carried by rivers to the shallow seas.”

  The Alvarezes are also suffering from too much of a good thing, for levels of iridium in the Danish sample are embarrassingly high. They calculate that average asteroidal iridium mixed with about 100 times its mass of earth material (the amount needed to conjure up the great dust cloud) should increase mean iridium levels in sediments by only about one-tenth the Danish amount. The Italian sample runs closer to expectations.

  Other problems may be in store. One uncertain study argues, from the fine stratigraphy of magnetic reversals, that the plankton line is not coincident with dinosaur extinctions—a small effect for the more leisurely concept of geologic suddenness (which might span many thousand years), but potentially fatal to the Alvarezes’ requirement for true simultaneity. I feel that the Alvarezes assume the proper attitude toward this report: they acknowledge it forthrightly, note its uncertainties, admit that its later confirmation would greatly weaken or destroy their hypothesis, and then predict—on the basis of their theory—that the report is both probably in error and greatly in need of more study.

  But I care little whether the asteroidal scenario itself is correct. The remarkable aspect of the Alvarezes’ work—the part that has produced buzzing excitement among my colleagues, rather than the ho-hum that generally accompanies yet another vain speculation—lies in their raw data on enhanced iridium at the very top of the Cretaceous. For the first time we now have the hope (indeed the expectation) that evidence for extraterrestrial causes of mass extinction might exist in the geologic record. The old paradox—that we must root against such plausible theories because we know no way to obtain evidence for them—has disappeared.

  When I started my career as a paleontologist, I used to argue that mass extinctions might be rip-roaring fun to discuss but relatively unimportant for the ultimate disposition of life and its history. I was then caught up in some common prejudices about inherent, stately progress as a hallmark of life’s history. The mass extinctions, I thought, might disrupt the process severely, setting the clock of progress far back. But time (to a geologist) is not a severe constraint; life would recover, moving on as before.

  I can’t think of any other idea I once held (after age five) that I now regard as so foolish (except the thought that the N.Y. Giants might catch the Brooklyn Dodgers in 1951, which they did—thanks again, Bobby Thomson). The history of life has some weak empirical tendencies, but it is not going anywhere intrinsically. Mass extinctions do not simply reset the clock; they create the pattern. They wipe out groups that might have prevailed for countless millenniums to come and create ecological opportunities for others that might never have gained a footing. And they do their damage largely without regard to perfection of adaptation (the most gorgeously designed photosynthetic plankter could not survive a great darkness, while some marginal competitor might squeak through and become a progenitor of the next dominant group).

  Who knows? Without the great Cretaceous extinction, dinosaurs might have rallied and still dominate the earth (they had already lived far longer than the sixty-five million years since their demise). Mammals might still be a small group of ratlike creatures casting about for an occasional bit of protein in a triceratopsian egg.
>
  Among the Cretaceous mammals that witnessed the great event, we know a single primate named Purgatorius. It may not have been the only member of our order, but there probably weren’t many of us back then. Suppose that Purgatorius hadn’t pulled through—and remember, it probably owed its continuation to luck or to adaptations not related to features we value in primates because we have capitalized upon them. Primates would not have reevolved. A giraffe on the plains might now be surveying the creation from on high and smugly regarding itself as the finest of all God’s creatures. Our current existence is an extended function of enormous improbabilities. We may owe our evolution, in large part, to the great Cretaceous dying that cleared a path, yet spared our ancestors’ lives to tread upon it. That asteroid may well have been the sine qua non of our existence.

  Postscript

  So much has been written and argued about the Alvarez hypothesis since this essay appeared in June 1980 that any attempt at complete (or even adequate) update would require a book in itself. I therefore decided to let the essay stand largely as it appeared; at least it may have some historical virtue as a comment upon the hypothesis as first presented. Since then, the primary gain in support for Alvarez has been the discovery of iridium spikes right at the Cretaceous—Tertiary boundaries in many other places beyond the original two spots mentioned in the essay. Some include very different environments (deep sea cores) and very different places (New Zealand, at the virtual antipode of the two original sites). All this augurs well for a hypothesis that requires a worldwide effect. The primary difficulty and source of debate has centered upon an issue prominently discussed in my essay (one that any professional paleontologist would—and did—raise). In their original article (see bibliography) Alvarez et al. strongly suggested that their asteroid might provide a complete explanation for the pattern of late Cretaceous extinctions. Yet the paleontological record indicates clearly (as critics have effectively established) that many groups meeting their end in the Cretaceous were in decline for millions of years before any late Cretaceous disaster could have struck the earth. In my view, this fact does not downplay the significance of a potential latest Cretaceous catastrophe; it does not even reduce such a catastrophe to the relatively insignificant status of coup de grâce. For, as I argue in the essay, these gradual declines—had no terminal catastrophe been superimposed upon them—would probably not have led to a major extinction that reset the pattern for life’s subsequent history, including our own evolution. We need a complex and synergistic theory for the Cretaceous extinction.