by Philip M. Dauber and Richard A. Muller

Helix Books, Addison-Wesley Publishing Company (1996)

CHAPTER EIGHT

Nemesis and Mass Extinctions

At crucial times in the fossil record, large numbers of apparently thriving species became extinct, all within a relatively short interval. Different creatures replaced them, living in similar climatic conditions, in the same geographical areas. These successor species were not necessarily better adapted or more "fit"_they just appeared later in evolutionary history. Of the many extinctions known, five stand out as more globally devastating than the others. Marking the boundaries of epochs in Earth's geological and biological history, these five great extinctions took place about 440, 365, 245, 210, and 65 million years ago. Of these, three extinctions have been linked with large Earth craters and three with boundary clays rich in iridium, suggesting an extraterrestial impact.

The last great extinction, the K-T catastrophe 65 million years ago, saw the end of about 40 percent of all animal genera (genera is the plural for genus, the category in biologists' class)fication system between family and species) and about two-thirds of animal species. Marine reptiles, including the huge long-necked plesiosaurs, the mosasaurs with paddles instead of legs, and the sharklike icthyosaurs, died out completely. Every single species of terrestrial dinosaur vanished; the birds, considered by many biologists to be a form of dinosaur, survived. Most flowering plant species endured and made a comeback after ferns took over the continents. (Such a takeover is exactly what happens when a forest burns today: Eventually young trees crowd out ferns, which cling to existence below the forest canopy.)

The fossil record does not tell clearly whether these great extinctions happened in a single day or over several million years. Geology is too uncertain a science for that, at least for now. Erosion makes the record difficult to read. Dating methods are too imprecise to allow paleontologists to compare reliably the fossil record at different sites around the world. Fossils of large animals like dinosaurs are rare. At one key site a 2or 3-meter gap lies between the most recent dinosaur skeleton and the iridium-rich clay layer. As Luis Alvarez and many others have pointed out, this gap does not necessarily mean that the dinosaurs died out before the impact. There might simply have been an interval during which no dinosaur was preserved well enough to become a fossil in the rather small fraction of the Earth's surface that has been searched. Since the average spacing between dinosaur skeletons is only about a meter, this possibility is statistically reasonable. In any event, we have no reason to expect a concentration of dinosaur fossils right at the K-T boundary.

In between the five great extinctions were many others_almost two dozen, according to paleontologist David Raup. In these lesser events, smaller percentages of species and genera became extinct. Two smaller extinctions, the ones dating from 35 million and 290 million years ago, have been tied to known craters.

Before the Alvarez work linked the K-T catastrophe with an impact, paleontologists had been debating the causes of mass extinctions for decades. Despite the general feeling that no one mechanism could possibly dominate in such a complicated process, climatic change, especially cooling and drying, was the most popular explanation. There have been many others, including the rises and falls of sea level, predation, epidemics, competition with other species, poisoning of ocean waters, changes in atmospheric chemistry, global volcanism, and comet or asteroid impact. Established species that are well distributed geographically are extremely difficult to kill off, Raup emphasizes in his studies of extinction; he came to the conclusion that some kind of extraordinary "first strike" was necessary before most of the proposed extinction mechanisms could have a credible chance to work. Was it possible that extraterrestrial impact has been the major cause of mass extinctions, maybe even the only cause?

University of Chicago paleontologist John Sepkoski had a special interest in the dates when particular types of fossil appear in the record and when they disappear. In 1982, after collecting fossil data for many years, he produced a massive compendium of 3,500 families and by 1984 a computerized index of 30,000 genera. Raup and Sepkoski suspected that this mass of data contained some simple patterns that could shed light on the mechanism of mass extinctions, but they had no particular mechanism in mind at that time. In 1977 paleontologist A1 Fisher had claimed to find a regular 32-million- year spacing between mass extinctions. Using a variety of computer methods to analyze the twelve biggest mass extinctions, Raup and Sepkowski found a regular spacing too, but at 26 million years rather than 32 million. Try as they might (and as good skeptical scientists must), they could not shake the regularity of the intervals from their calculations.

A dozen or more groups have reanalyzed the Raup and Sepkowski data to see if mass extinctions do have periodicity. According to David Raup, "The results are mixed: about half support the 26-million-year periodicity (with minor revisions of the period in some cases), and about half find no convincing evidence for cycles of any duration." Raup himself still believes the periodicity is real, but most paleontologists do not. Another, more subtle objection is that the periodicity in the fossil data is real but is the result of a characteristic "healing time" that life needs in order to recover after any deadly impact rather than a regular period of the impacts themselves.

The clocklike spacing between extinctions was nonetheless intriguing. One of the present authors, Richard Muller, caught wind of the Raup and Sepkowski finding before its publication. Muller came up with a possible explanation_our Sun may have had a little companion star orbiting around it with a 26-million-year-period. Most stars, after all, are part of binary systems. The two stars closest to Earth, Alpha Centauri and Proxima Centauri, orbit each other. If our Sun's hypothetical companion star approached the inner solar system every 26 million years, it might knock lots of asteroids out of their usual orbits. One or more of these asteroids could hit the Earth and cause extinctions. Not only did this explanation provide a plausible clock in the sky, it had an important side benefit: The asteroids could come in bunches. This would answer paleontologists who kept objecting to the impact theory on the grounds that the dinosaurs had died off over hundreds of thousands if not millions of years, rather than all at once. Okay, maybe it took several crashes to do in the dinosaurs, astronomers would be able to reply.

Unfortunately, this first explanation of the periodicity of mass extinctions had a fatal flaw: An orbit that brought a companion star close enough to the Sun to enable it to fling asteroids from orbit would be very elongated and unstable. The pull of passing stars would change the orbit so much that on the next pass it would come nowhere near the inner solar system. A changing orbit could not explain periodicity.

Muller, teaming up with astronomers Marc Davis and Piet Hut, soon came up with a workable revision of the theory. Suppose the companion star were on a much less elongated orbit, one shaped roughly like an egg. Its maximum distance from the Sun would be about 3 light-years and its minimum distance about half a light-year. (Half a light-year may not sound like much, but it is a distance 160 times greater than the orbital distance of Pluto from the Sun.) This more circular orbit would be much more stable and could produce periodic impacts.

Every 26 million years, the companion would pass through the Oort comet clouds. There, just as Oort had theorized for randomly passing stars, it would destabilize the orbits of billions of comets. Some would gain energy and speed and be ejected from the solar system. Others would lose energy and begin a long fall toward the Sun. For every billion comets dislodged, the team calculated, about a million would cross Earth's orbit. Of these, about two would hit the Earth. The numbers came out right: The bombardment would last about a million years, and during that time a new comet would be visible from Earth about every three days. Very few of these would hit. For each complete cycle, sometimes one comet would hit Earth, and sometimes two, three, four or five. Sometimes, by pure chance, none would hit.

Muller suggested that the companion star be named Nemesis, after the Greek goddess who made sure no earthly beings challenged the dominance of the gods. Before the startling new hypothesis could be published, however, a serious question had to be addressed: Would the companion star's orbit be stable, or would it too be disrupted by passing larger stars? Piet Hut published a calculation that showed that the duration of the present Nemesis orbit (its "lifetime"), if it existed, would be about 1 billion years. This means that in the next billion years, there's about a fifty-fifty chance of a passing star kicking Nemesis out of its liaison with the Sun, thus ending its reign of terror. The billion-year figure is based on the present size of the orbit of Nemesis. According to Hut's calculation, the orbit of Nemesis has grown steadily over the 5-billion-year lifetime of the solar system. The companion star was once much closer in, with a correspondingly shorter orbital period. When Nemesis was originally formed along with the Sun and planets, its orbit then lasted about 5 billion years. Encounters with passing stars, on average, have tended to increase the energy of Nemesis so that its orbit lengthens. This is analogous to what Nemesis does to comets in the Oort cloud: It boosts their energy so that more of them leave the solar system than the number that lose energy and fall inward.

If the Nemesis theory is correct, geologist Walter Alvarez quickly realized, there should be evidence for it in the cratering record on Earth, such as a regular pattern in the impact dates. He and Muller began searching for periodicity in the ages of the impact craters on Earth. The first plots they studied were disappointing: No pattern jumped out at them. But many craters are very poorly dated. Uncertainty in their ages could easily blur a 26-million-year period. Alvarez's solution was simple: Ignore the craters that don't have accurate dates. When the list of nearly a hundred craters was culled down to two dozen, something exciting happened. There were three or four obvious pairs spaced by 30 million years or so, and these included some of the largest craters. When only the larger craters were plotted, clumps of them seemed to be spaced by 26 to 30 million years. The average interval was 28 million years. A detailed statistical analysis came next. Fourier analysis, a powerful mathematical technique that finds periodicity in data, produced a strong peak at a period of 28.4 million years, towering above any other possible regularity in the data. When the computer was called upon to calculate crater ages distributed randomly, the Fourier program found a comparably large peak only every few hundred tries. That is enough statistical sign)ficance to be interesting, but it is not overwhelming evidence for a new theory.

Claimed "discoveries" in science have a long history of appearing to be moderately convincing on a statistical basis but going away later when more data come in. In the long run, it doesn't help a scientist's reputation to be involved in making such a claim, even if the discovery paper catapults him or her to overnight fame. In this case, Walter Alvarez and Rich Muller were well enough known already that they probably had more to lose than to gain from publishing the Nemesis theory. Luis Alvarez, for his part, had seen plenty of these scenarios, and he certainly wanted to protect his son and Muller from a possible mistake. He had been alternating between extreme enthusiasm for and skepticism about the Nemesis theory, but finally he tried vigorously to discredit the periodicity in the crater data by showing that the data were not statistically sign)ficant or were otherwise flawed. After several intense weeks of give and take with his son and Muller, Luie relented, and his skepticism abated. The Nemesis paper was sent to the journal Nature for publication.

Louisiana astronomers Daniel Whitmire and Albert Jackson had independently come up with a similar theory to explain the apparent periodicity of mass extinctions_comet showers triggered by a companion star_but they hypothesized an extremely eccentric (and, it turned out, unstable) orbit. They sent their paper to Nature too. Whitmire also advanced the idea of a Planet X orbiting beyond Pluto. Such a planet, he showed, could perturb the inner part of the comet cloud and cause comet showers. But Planet X would spread the comet showers over many millions of years, it turned out, in contradiction to the mass extinction data. Another theory focused on the bobbing motion of the Sun in and out of the galactic plane. Well known to have a period of about 33 million years, this motion could cause periodic perturbations of comets in the Oort cloud, due to the concentration of stars in the plane and the lack of stars out of it. Unfortunately for this theory, there is no correlation between the actual crossings of the galactic plane by the Sun and the ages of the mass extinctions. Also, the bobbing of the sun is too small to give the necessary effect.

One could hardly imagine a more sensational and controversial scientific hypothesis than an invisible killer star orbiting the Sun flinging deadly missiles at Earth. Scientific journals in 1984 were filled with earnest debate, but some of the utterances of usually polite scientists were, to put it diplomatically, less than temperate. The explosion of media interest was incredible. Time magazine put the story on its cover. There were TV documentaries, endless TV interviews with the scientists involved, even editorials in The New York Times. In 1985, one of these, entitled "Miscasting the Dinosaur's Horoscope," concluded

Terrestrial events, like volcanic activity or changes in climate or sea level, arethe most immediate possible causes of mass extinctions. Astronomers shouldleave to astrologers the task of seeking the cause of earthly events in the
stars.

Like the tides and seasons? Walter Alvarez and Richard Muller replied in a letter to the Times.

You say, "Complex events seldom have simple explanations." The whole history of physics contradicts you. You suggest "Astronomers should leave to astrologers the task of seeking the causes of earthly events in the stars." May
wesuggest it might be best if editors left to scientists the task of adjudicating scientific questions.

The famous biologist Stephen J. Gould lampooned the Times with this piece of parody, supposedly taken from an Italian newspaper of 1663:

Now that Signor Galileo, albeit under slight inducement, has renounced his heretical belief in the earth's motion, perhaps students of physics will return to the practical problems of armaments and navigation and leave the solution of cosmological problems to those learned in the infallible sacred texts.

Carl Sagan found the Nemesis theory

serious and respectable, if highly speculative, science, because the principal idea is testable: You find the star and examine its orbital properties.

Sagan wrote his own letter to The New York Times defending the Nemesis theory. For several years the debate raged inconclusively.

Today, the concept that extraterrestrial impacts cause devastating catastrophes on Earth is well accepted. Also extremely persuasive is the evidence linking the K-T mass extinction with the impact at Chicxulub crater. Many people, including astronomers who should know better, believe that the Nemesis theory has been disproved, that the orbit has been shown to be unstable. Of course the orbit is unstable, with an expected lifetime of about a billion years, as shown in the orginal Nemesis paper and as subsequently verified in detail by Hut. Since the solar system is 5 billion years old, many people think that a billion- year orbit could not still be in existence; the orbit would have lasted only 1 billion years into the life of the solar system. But they are confusing the present lifetime of the orbit with its past duration. The lifetime of the Nemesis orbit has been slowly decreasing since the formation of the solar system. Five billion years ago, Piet Hut showed, the expected duration of the Nemesis orbit was 6 billion years. According to the theory, Nemesis lhas only 1 billion of these years left.

There may be good reason to question the Nemesis theory, but the obvious question is, Why hasn't Nemesis itself ever been seen? At a distance of three light-years from us, it would be the closest star, more than a light-year closer than the Centauri pair. Basically the answer is that Nemesis, if it exists, is too small and too dim. (We are assuming that Nemesis is an ordinary red dwarf, like most visible stars much smaller than the Sun.) In order to have enough gravitational kick to launch comets at Earth, Nemesis would have to have a mass of at least onetwentieth that of the Sun. If its mass were as large as one-third that of the Sun, it would have been seen_in fact, it would be brighter than Proxima Centauri, and its extreme nearness to us would be well known. But if its mass were much less than a third of solar mass and its brightness correspondingly reduced, Nemesis would be an obscure dim star that looked just like many intrinsically brighter stars much farther away. Its proximity would probably have been missed by standard astronomical surveys. Until the Nemesis theory, astronomers had little reason to make the measurements necessary to reveal the closeness of a dim star.

To convince the world (and ourselves) that Nemesis is real would require discovery of the star itself. Hunting for Nemesis, however, is like looking for the proverbial needle in a haystack. Astronomers measure the distance to nearby stars using a method based on the phenomenon of parallax. To grasp this concept, extend your finger in front of you and close one eye. Note the position of your finger relative to some object in the background, such as a picture on the wall. Then switch eyes. Your finger seems to jump. It has a different position with respect to the fixed background as seen by your other eye. That's parallax. Astronomers observe a star at three-month or six- month intervals and carefully measure its position relative to other stars, especially ones known to be very far away. What they are doing is observing the star from different parts of Earth's orbit around the Sun. The position of a nearby star is measurably different when seen from different sides of Earth's orbit. Finding Nemesis requires surveying thousands of stars at different times of the year, then comparing the images with great precision. Using an automated telescope to survey the northern hemisphere, the Berkeley group has eliminated more than half of 3,100 candidate red dwarf stars. On every clear night about ten stars can be checked off. The search will go on until no candidates are left_or Nemesis is found.

Nemesis may not be a red dwarf at all. It could be an exotic object like a black hole, a neutron star, or a brown dwarf. Gravitationally, its action in dislodging comets while circling the Sun would be the same as a red dwarf's. But it would be next to impossible to detect, and we have no reason to suppose that such strange objects are common in this part of the Milky Way galaxy.

Nemesis may not exist at all, and impacts may not even be periodic. Could mass extinctions still be caused by comets and asteroids? Paleontologist Raup thinks they could. As we mentioned earlier, he thinks a "first strike" must reduce the geographical range of a healthy species before it can become vulnerable to extinction. Comet showers require neither Nemesis nor periodicity. Any passing star that perturbed the orbits of lots of comets could produce a deadly shower.

Nor is the geological record clear on the link between impacts and extinctions. Of seven boundary layers in the Earth where iridium is in excess, Raup finds that four are associated with important mass extinctions, but the others are questionable. And of fourteen craters at least 32 kilometers in diameter and younger than 500 million years, five (including Chicxulub) match the ages of mass extinctions. This information is intriguing but inconclusive. The greatest mass extinction of all, the one 245 million years ago, has not been linked to any impact. How any major impact could fail to produce extinctions is unclear, given the enormous release of energy and long list of atmospheric horrors that accompany it. But what is a "major" impact? We don't know the impact energy threshold above which mass extinction is inevitable_if we tried to guess, we could be wrong by a factor of ten, even a hundred. There are other variables, too, that affects an impact: the type of rock struck helps determine the nature of the dust cloud and the intensity of the killing acid rain that accompanies it. The very largest known crater, Chicxulub (170 kilometers), is definitely associated with mass extinctions, and so is the second largest, Manicouagan (100 kilometers) in Quebec, which is linked to the 208-million-year-old great extinction between the Triassic and Jurassic eras. The craters not yet matched with extinctions are in the 50-kilometer range; the energy needed to make them is at least ten times less than that required for Chicxulub.

If the Nemesis theory is correct, where do we stand now in the extinction cycle? The most recent mass extinction occurred about 14 million years ago. If Nemesis were to blame, it must have made a pass through the Oort cloud about 14 million years ago. Now it would be near its farthest distance from the Sun, destined to return to the cloud in about 12 million years. So for now we are safe_at least from Nemesis. But the Nemesis theory does not claim that all large impacts are due to comets dislodged by Nemesis, or even due to comets at all. Some may be caused by rogue asteroids. Don't get too comfortable! For although we humans owe our "successful" evolution to an impact, we might someday owe our doom to one too. What can we do to guard against this frightening possibility?