—DAVID M. RAUP AND J. JOHN SEPKOSKI, JR., “PERIODICITY OF
EXTINCTIONS IN THE GEOLOGIC PAST,” PROCEEDINGS OF THE
NATIONAL ACADEMY OF SCIENCES OF THE U.S.A., VOLUME 81, P. 801, 1984
Every 26 million years or so, like clockwork it seemed, plants and animals drop dead all over the planet. At the family, genus, and species levels, they concluded, a major fraction of all life on Earth becomes extinct at apparently regular intervals. Extinction. Scientists tend to talk about it with detachment, but there is something unnerving about these countless lives being snuffed out, their ancestral lines rendered meaningless, by some regular motion of the Grim Reaper’s scythe. Naturally, we wonder if we may be next.
Paleontologists have for decades invoked a wild variety of explanations for the mass extinctions, but none of the proposed mechanisms had a period built in. The Earth does not seem to have any 26- or 28- or 30-million-year-long cycles—not in volcanos, not in plate tectonics, not in the weather. Periods this long fall in the domain of astronomy.
The iridium in the Cretaceous boundary has made the idea of periodic (or at least episodic) mass extinctions, visited on the Earth from space, more palatable; but for many scientists, and others, it remains a bitter pill to swallow. Even if for the moment we grant that small worlds fall out of the sky like clockwork, shouldn’t this fact be evident from the ages of craters on Earth? The craters are there—little recent ones, not yet eroded away, and larger and older craters, the biggest being more than 100 kilometers across. Still larger craters must have been made, of course, but that was so long ago in Earth’s history that the slowly crashing and diving plates that make up the Earth’s crust would have destroyed them long ago. Thus newly motivated, several teams of scientists have examined the ages of the surviving craters, and have seemingly found—to simultaneous satisfaction and amazement—that craters tend to have been excavated, by the impacts of objects from space, around every 28 million years, very close to the inferred period of the mass extinctions. What’s more, the events are reported by some to have been in phase: mass extinctions tend to happen when large craters are formed, the two events presumably caused by the same impacting bodies.
Glass
Certain sedimentary layers of the Earth are strewn with glassy, geometrically smooth inclusions, ranging from a few centimeters in size to submicroscopic They are called tektites. That tektites are produced when comets hit the Earth was first proposed by the American chemist Harold Urey in 1957, and—while not definitively decided—the hypothesis has stood the test of time. The comet strikes the Earth, digs out a large crater, and melts the underlying ground in the process; droplets of silicates are flung over huge distances, freezing in the process and producing streamlined, sometimes teardrop-shaped, forms. The tektites show signs that they themselves were cratered during these violent events. It has been claimed and disputed that layers of sediment containing the so-called microtektite horizon are associated with the Eocene extinctions. One of the leading experts on these tiny glass forms is named, appropriately enough, Billy Glass. He is at the University of Delaware.
The correlation is not one-to-one. In sediments older than the Cretaceous, dating of these boundaries has an uncertainty of many million years or more; this is one reason—there are several—why a precise concordance between crater ages and extinction times could not exist. Some cosmic objects, as probably at the end of the Cretaceous, must have crashed into the oceans, leaving no visible crater. And, as we would expect, there are random craters that are not part of any 28-million-year cycle.
Since these heady announcements were made in the scientific literature, a number of scientists have gone back to the original data to check the reliability of the conclusions drawn. How complete is the list of craters? How accurately determined are the times of impact? How do you define a mass extinction? How reliable are the dates of cratering or mass extinctions? Suppose cratering and extinction events happened entirely randomly; how likely is it that you would by chance find spurious evidence of periodicity? The results of this reanalysis are still not yet fully in, but in the heated debates that have ensued it is clear that the foundations of several sciences are being constructively reassessed. There are those who believe in an in-phase periodicity of cratering and mass extinctions. There are those who believe that cratering may be periodic but mass extinctions not, or the reverse. And there are many scientists who hold that, within the present uncertainties, there is no compelling evidence for periodicities either in cratering or in mass extinctions.
… The evidence for [periodic mass extinctions] is strongly contingent on arbitrary decisions concerning the absolute dating of stratigraphical boundaries, the culling of the database and the definition of what is mass extinction as opposed to background extinction. This evidence becomes insufficient under other plausible geological timescales and other acceptable definitions of mass extinction. Analysis of the non-culled database shows that the reliability of identification of mass extinctions and their timing is at present extremely limited. It also suggests that the apparent periodicity of mass extinctions results from stochastic [random] processes.
—ANTONI HOFFMAN, “PATTERNS OF FAMILY EXTINCTION DEPEND ON
DEFINITION AND GEOLOGICAL TIMESCALE,” NATURE, VOLUME 315, P. 659, 1985
If this last group of scientists is right, then the remainder of this chapter is only a flight of fancy. But if there are such periodicities, a stunning connection between life down here on Earth and events up there in the sky has been uncovered. In that case, somewhere a cosmic doomsday clock is ticking away even now. Fortunately, we are today midway between mass extinctions. The next bombardment is not scheduled, if the periodicity is correctly deduced, for another 15 million years. Our job is to avoid becoming one of the background extinctions.
But how could any object in space know when its time had come to hit the Earth? What cosmic machinery could serve as a doomsday clock? Take, for example, the present population of short-period comets, or the hundreds of asteroids that cross Earth’s orbit. They will strike the Earth from time to time, of course, and contribute to the background rate of cratering. Indeed much of the background iridium in rocks, unassociated with biological catastrophes, may also come from comets—the steady infall from meteor streams and the zodiacal dust. But this will be a constant pitter-patter; these worlds do not save themselves up for mass assaults on the Earth every 30 million years—or, for that matter, on any other time scale. Periodic impacts must be caused by other bodies, more distant.
By the mid-1980s two quite different proposals were in contention, each purporting to show how a cosmic metronome might beat out extinctions at 30-million-year intervals. Neither is fully satisfying; both have deficiencies. It is hard even to state the two hypotheses without seeming a little lurid, and the more sensational recountings in the popular press have in the minds of some scientists confirmed their initial misgivings. But together, they provide a good example of science in transition: A powerful conclusion is drawn from disputed data. But the conclusion poses an enigma. Two hypotheses attempt to explain the enigma in different ways. If science is to be served, the hypotheses should make different predictions about what you will find if you perform a particular sort of new experiment. Experimental confirmation of quantitative predictions is to science what the fulfillment of prophecy is to religion.
In one of the two hypotheses, life on Earth is periodically extinguished because the Milky Way Galaxy has the shape first understood by Wright and, especially, Kant (Chapter 4). Our Galaxy is a thin disk that contains the spiral arms and that rotates about a bulging core of stars and dust. It is at the center of the Milky Way that huge numbers of stars are concentrated; it is, in brilliance, mass, position, and explosive violence, the downtown of the Galaxy. Fortunately for us, we live nowhere near it, nor even in the suburbs. Our home is in an obscure galactic countryside, where stars like the Sun take 250 million years to go once around that distant hub.
But in addi
tion to revolving around the galactic center, the Sun has another motion: it is bobbing up and down, each time passing through the imaginary plane of symmetry that cuts through the galactic center. When at its maximum excursion, about 230 light-years above the plane, the Sun is gravitationally attracted by the gas and dust and stars beneath, slowly reverses direction, and falls back. But the galactic plane is an imaginary, not a real, surface, and when the Sun arrives there, it finds it has a sizable velocity and nothing to stop it. Accordingly, it plummets through and out the other side, slowing because of the gravity of the dust and stars it has left behind—until it is some 230 light-years on the other side of the galactic plane, stops, and then falls back again.
Since the Sun is traveling through a vacuum more nearly perfect than anything we know, there is no friction to retard its motion. Like a weight on a perfectly elastic spring, the Sun caroms up and down forever. The oscillating Sun is effectively a perpetual motion machine. As long as a great deal more matter doesn’t move into or out of our neighborhood in space, this galactic bounce will continue forever. The period of oscillation depends only on how much mass there is in the space near the Sun, and this seems to be well-measured by astronomers. Also, the Sun’s motion can be measured by examining the nearby stars with respect to which our system is moving. In both these ways it is found that the period from one crossing of the galactic plane to the next for the Sun and its entourage of planets and comets is about 33 million years. In fact, all the nearby stars are bobbing up and down, crossing the galactic plane once each 30 million years or so. This period is fundamental for millions of suns. You look at the ages of fossils and craters on Earth, and you find a time scale of almost 30 million years. You look at the Sun bobbing in the Galaxy and you find another time scale of roughly 30 million years. It’s hard not to think that the two periods might be related, that the oscillation causes the extinctions.
But how could the Sun’s motion instruct comets to collide regularly with the Earth? There are giant molecular clouds spread out in dribs and drabs in our part of the Milky Way. They do not all move at the same speed as the Sun, and every now and then the Sun is bound to run into one of them. Extending over a much larger volume than the solar system, they are much more massive than the solar system. Passing by or through such a cloud would gravitationally produce a great flurry among the comets of the Oort Cloud, and pry loose a shower of comets to descend on the Earth and its neighboring worlds.
The motion of the Sun in the Galaxy. The Sun orbits the massive, bright, bulging core of the Milky Way Galaxy once every quarter billion years. As the Sun slowly circumnavigates the center of the Galaxy it bobs up and down at a faster pace, making a bounce every 60 millions years or so, and therefore crossing the plane of symmetry of the Galaxy roughly once every 30 million years. Does this 30-million-year period somehow trigger mass extinctions on the Earth? Diagram by Jon Lomberg/BPS.
Comet Showers
Before a periodicity of extinctions and cratering was proposed, J. G. Hills of the Los Alamos National Laboratory suggested* that a passing star—not a companion star—in the inner Oort Cloud could stir up a storm of comets:
The observed comet cloud may be only the outer halo of a much more massive comet cloud whose center of mass is well inside the observed inner boundary of the Oort Cloud.
Hills then goes on to deduce that a star passing only 3,000 Astronomical Units from the Sun would produce a cometary shower from the inner Oort Cloud that would generate one new comet per hour in the vicinity of the Earth. This would have many consequences, but the first one that Hills, an astronomer, mentions is this:
Such a high comet flux would be a major nuisance to astronomical observers engaged in research on low-light objects!
This is another of the rare astronomical exclamation points. Hills then continues:
The integrated comet flux from such a shower can be great enough that several comets will actually hit the Earth during the shower. This may show up in the geological record.
For this metronome to work, though, the clouds must be concentrated in one particular place—the galactic plane, probably. Then, every 30 million years, the Sun plunges through the plane (alternately from above and from below), a shower of comets is launched toward the inner solar system, one or more of them strikes the Earth, the lights go out and the temperatures drop. If the interstellar clouds lived in the galactic plane, the oscillating Sun explanation of periodic mass extinctions would seem very promising.
However, for the whole 230 light-years of the Sun’s maximum excursion from the galactic plane, the clouds are distributed essentially at random. There is no more reason to bump into one in the galactic plane than there is 230 light-years above or below. Indeed, in the present epoch the solar system is only about 25 light-years above (north of) the plane of the Galaxy, and there is no giant interstellar cloud at our doorstep. The extinctions should, therefore, occur at random. The roughly 30-million-year period of galactic plane-crossing by the Sun does not seem to translate into a roughly 30-million-year cycle of mass extinctions on Earth (even ignoring the differences among 26-, 28- and 30-million-year periods). Maybe there’s some other way in which the Sun’s periodic bobbing in and out of the galactic plane can stir the comets up and trigger global catastrophes on Earth. Maybe there’s a flattened population of small primordial black holes in the Galaxy, through which the Oort Cloud runs as the Sun bobs. But if there is, no one has been able to find it so far,* and without something of the sort, this hypothesis, tantalizing as it is, remains unsatisfactory. What is the alternative?
Most of the stars in the sky are members of double or multiple star systems. In a typical binary system, two stars separated by several Astronomical Units are doing a stately gravitational fandango. Often the stars are more widely separated. In some instances we see two stars gravitationally bound to each other, but separated by 10,000 A.U. At least 15 percent of the stars in the sky seem to have a companion star at this distance. The nearest star system to the Sun—Alpha Centauri, 4.3 light-years away—is a double star with a third sun, a distant dim companion called Proxima Centauri, at 10,000 A.U. from the two bright stars. Often the companion star is very faint, suggesting that there may be many still undiscovered widely separated binaries. It is possible that most of the stars in the Galaxy are so dim that astronomers call them brown or black dwarfs. Most distant companions might be of this sort.
The solar system seems to be an exception. We do not know of any companion to the Sun. But if we were not an exception, if the Sun had an invisible companion star in a very specific orbit, then the extinction clockwork again might be understood. Suppose there was a companion star on a very long elliptical orbit—so that, on average, it was 90,000 A.U. away, 1.4 light-years. But once each orbit it comes much closer to the Sun, maybe 10,000 A.U. or even a little closer. This would bring the star into the inner part of the Oort Cloud, where comets are not ordinarily jostled by passing stars. With such an orbit, once every 30 million years the companion would plow through the denser parts of the Oort Cloud, and shower the Earth and neighboring worlds with comets.
The year of this star would be 30 million of ours. Ten times around the Sun would carry it from the Permian, when there were dragons with sails on their backs, to now, when humans reconnoiter the planets and menace the Earth. At present, it would be near aphelion, the farthest point in its orbit from the Sun. Fifteen million years from now, it would run its next rampage among the comets. During its plunge through the inner Oort Cloud, the companion star would spray a billion comets into the inner solar system. But because they would be carried on slightly different trajectories, the comets would not all arrive at the same instant. Rather, they would be spread out over a million years or more. Thus, the companion star hypothesis predicts that the background cratering and extinction rates will be periodically interrupted by intervals lasting as long as millions of years, in which the Earth is hit by several, perhaps several dozens of comets. This would explain why mass e
xtinctions do not occur instantaneously but sometimes over a period as great as millions of years.
The hypothetical companion star’s period around the Sun cannot be exactly 30 million years; indeed, it cannot be exactly anything: gravitational perturbations by stars, interstellar clouds, and the massive core of the Galaxy will tug it first one way and then another, altering its orbital period. Eventually, any such orbit will be changed greatly. A companion star in the present orbit could survive at most a billion years before it is stripped away by passing stars and the general gravitational pull of the Milky Way. This is long enough to account for periodic extinctions and impact craters back to the Permian—but, before that, the companion star would have had a very different orbit. If a companion star began much closer to the planets, it would have produced major and almost continuous cometary showers early in the history of the solar system. However, such a greatly enhanced comet flux early in solar system history is expected anyway, as comets were cleared out from the region around Uranus and Neptune—being propelled both outward away from the Sun to populate the outer Oort Cloud, and inward to crater the terrestrial planets (Chapter 12).
But before we seriously consider such extravagant possibilities, the hypothesis must face one awkward fact: There is, to date, no evidence whatever of a companion star. It need not be very bright or very massive; a star much smaller and dimmer than the Sun would suffice, even brown or black dwarfs—giant planet-like bodies insufficiently massive to generate hydrogen fusion at their cores and burst forth with stellar light. It is conceivable that a companion already exists in one of the catalogues of dim stars without anyone noting something anomalous—an enormous apparent motion of the star each year against the background of more distant stars (parallax, Chapter 2). There has been at least one major search for dark, cool stars, designed specifically to catch a companion star, if one exists. If one had been found, and in something like the right orbit, few would have doubted that it was the principal cause of periodic mass extinctions on Earth.* Now, it remains a provocative but unproved hypothesis.