Nemesis Strikes Back: Lisa Randall's Dark Matter and the Dinosaurs

Derek Turner writes . . .

Where Physics Meets Paleontology

Lisa Randall is not the first physicist to weigh in on the extinction of the dinosaurs. It was a physicist, Luis Alvarez (working with his son Walter, a geologist, as well as two chemists, Frank Asaro and Helen Vaughn Michel) who first presented evidence that a meteoroid strike brought the Cretaceous to an end.[1] Randall’s new book, which has been positively reviewed (here, here,  and here) follows in this proud tradition. A connection between dark matter and mass extinction would indeed be astounding. But I want to raise some questions about Randall’s explanatory story that I hope will resonate with paleontologists. I wonder, in particular, about her decision to focus on the geological record of impact craters rather than the fossil record.

First, a qualification: I’m not really qualified to write a review of a book that’s largely about the physics of dark matter. But the basic gist of Dark Matter and the Dinosaurs is easy enough to understand. Lisa Randall’s goal is to develop a story about the causes of mass extinction which she admits up front is rather speculative, but which nevertheless has some basis in current work on the physics of dark matter, including some of her own research. Indeed her book is a lovely example of virtuous speculation in science (see Adrian's discussion of speculation here).


Did Dark Matter Finish off the Dinosaurs?

Much of the stuff in the universe is dark matter, which doesn’t interact with light, so is, well, “dark,” invisible. It does, however, have weak gravitational interactions with ordinary matter (those familiar electrons, protons, and neutrons). How do we know that dark matter is real? Positing its existence can help explain some otherwise puzzling astronomical phenomena. It also helps to fill in some otherwise puzzling gaps in cosmologists’ understanding of the early history of the universe. Efforts to detect it indirectly are underway. It’s also not clear exactly what dark matter is, but physicists are working on that, too. Whatever it is, dark matter is probably passing through our bodies all the time without interacting with the stuff we are made of.

So far, so good. Now, consider that our entire solar system is in motion. The Milky Way galaxy forms a disk, with a bulge in the center. Our solar system is some ways out from the center and moving slowly around. A single lap around the galaxy takes a couple hundred million years, and we have only made a few. But as our solar system does these laps, it bobs up and down. Sometimes we’re closer to the “top” of the galactic disk, sometimes closer to the “bottom.” As the solar system bobs up and down, it passes through a narrower disk of dark matter that is inside the Milky Way. Randall's own original research, suggesting that there might be different kinds of dark matter that interact with ordinary matter in different ways, helps explain the formation of this dark matter disk. The solar system passes through the dark matter disk once every 30 odd million years. All that dark matter has just enough gravitational oomph to influence the solar system’s less stable outer fringe, especially the Oort cloud, which astronomers think to be the source of most of the comets that we see. The gravitational force of the dark matter disk dislodges stuff from the Oort cloud, sending comets zooming toward the inner parts of the solar system, where some of them impact Earth. Randall speculates that the meteoroid that caused the Chicxulub crater in Mexico 66 mya—the one responsible for the end-Cretaceous mass extinction event—was just one of these.



Perhaps one could say that dark matter is our Nemesis, since Randall’s story has a lot in common with the Nemesis hypothesis, which scientists entertained for a while in the 1980s, following David Raup and Jack Sepkoski’s claim that mass extinction events in Earth’s history exhibit a 26 million year periodicity.[2] In their original paper, Raup and Sepkoski hinted at the need for some sort of extra-terrestrial explanation, largely because they couldn’t see how any biological mechanism could explain the extinction periodicity. Physicists soon followed up with a story about a heretofore undetected companion star to our sun, which they called “Nemesis,” and whose approach to our solar system would cause trouble in the Oort cloud every 26 million years or so.[3] As Randall explains (p. 260), Nemesis never panned out. If it were out there, astronomers would have observed it by now. Raup went on to write this book about the episode.


The Fossil Record vs. the Crater Record

It’s possible to read Randall’s book as something of a vindication of Raup and Sepkoski’s original idea that we need an extra-biological cause for extinction periodicity. But she turns from paleontological research on extinction periodicity (which is admittedly messy, complicated, and controversial), and focuses instead on the record of big (20 km+ in diameter) impact craters over the last 250 million years. She seems less interested in efforts by paleontologists to identify a biological signal in the fossil record—as she herself says, she places extinction “on the back burner” (p. 353)—and focuses instead on the geological signal of impact periodicity. The following diagram may help to clarify her explanatory story:

That is, Randall zeroes in on the relationship between the solar system’s journey and the periodic meteoroid strikes. Her approach suggests that the solar system’s regular encounters with the dark matter disk are the ultimate cause of the extinction periodicity in the fossil record. 

Why does Randall shift the focus from the fossil record—which is what Raup and Sepkoski were studying—to the impact crater record? Here’s what she says:

Meteoroid hits are challenging enough to investigate. Coupling them with uncertainties about extinction events is bound to go down a convoluted rabbit hole of trouble.

    Because of these uncertainties—apart from the lone well-established meteoroid/K-Pg connection—the rest of this book will shy away from further speculations about extinctions, intriguing as they may be” (p.243).

In one respect, her approach makes very good sense: in the chain of causation, impact craters stand between comets dislodged by dark matter and any resulting extinction pulses. So the crater record will give us more direct evidence of the action of dark matter than the fossil record will. Still, it’s hard (for me, anyway) to shake the feeling that mass extinction is what we really wanted to know about in the first place. To be fair to Randall, there is one case where the connection between the meteoroid impact and the extinction pulse is fairly well understood—namely, the end-Cretaceous extinction. But one case isn’t really enough, given that the larger explanatory story is about patterns: how is the pattern in the impact crater record related to the pattern in the fossil record?

Arizona's Meteor Crater, only about 50,000 years old, the result of a much smaller meteoroid than the one that finished off the dinosaurs.

Arizona's Meteor Crater, only about 50,000 years old, the result of a much smaller meteoroid than the one that finished off the dinosaurs.

Notice how Randall’s shift of focus from the fossil record to the crater record leaves some empirical questions unaddressed. What if the periodicity in the two data sets does not line up just right? In other words, what if there are impact craters that don’t correspond with extinction pulses in the fossil record? This result would deepen the mystery: why do big meteoroid impacts sometimes trigger mass extinctions, sometimes not? Conversely, extinction pulses without impact craters might suggest either (a) that we haven’t found all the craters, which would raise questions about the completeness of Randall’s data set or (b) that meteoroid impacts just aren’t the whole story about extinction patterns.

Even small mismatches between the crater record and the fossil record might raise questions about the case that we think we understand pretty well—the K/P-g mass extinction of 66 mya. (See Patrick Forber and Eric Griffith's paper for a philosophical perspective on that event.) Nobody seriously doubts that the meteoroid that left the Chicxulub crater caused a lot of biological drama and trauma. But the extinction patterns in the fossil record are messy and complicated, with some evidence of decline in some groups before the impact. It’s difficult to get precise about when different groups disappeared, and there were other potential contributing causes, such as the volcanic activity that created India’s Deccan traps. I myself don’t really know the answers to these empirical questions, of course – my suggestion is only that much of the scientific fun (and challenge) lies with the issues that Randall sidesteps, especially the effort to integrate what we know about the fossil record with what we know about the record of impact craters. And until that integration takes place, we should be cautious about giving too much destructive credit to dark matter. That is, for the universe to be as interconnected as Randall suggests, we’d better connect the crater record to the fossil record.

Other scientists have been working on this problem. See, for example, this recent paper from Michael Rampino and Ken Caldeira, who argue that the two records line up quite well, though there do seem to be some peaks in the impact record that do not correspond to any mass extinctions. That's an important detail, because it might suggest that impacts don't always cause mass extinctions.  Another detail is that Rampino and Caldeira find peaks in the crater record at 18.4 and 25.8 million years, whereas Randall works with something closer to a 35 million year periodicity in the crater record. 


How Complete is the Crater Record?

One possible justification for shifting for focus from extinction periodicity in the fossil record to impact crater periodicity is that the latter data set is more complete – and thus could be a better record than the other. Randall hints at this argument (p. 243), but doesn’t really spell it out. And she herself refers to “the paucity of the crater record” (p. 162). In her chapter on “Shock and Awe,” she gives a vivid description of how impact craters are formed on land, but as she also notes, oceans cover most of the planet. What happens when a meteoroid hits the ocean? What signature(s) would it leave? As she points out, tectonic subduction means that the seafloor is relatively younger—mostly less than 200 million years—because portions of it are steadily getting conveyor-belted back down into the mantle. This suggests that incompleteness is a big problem in the crater record: How often should we expect meteors to hit the oceans? By contrast, the fossil record for many kinds of marine organisms is pretty good. So while it could turn out that the impact crater record is a better data set than the fossil record, I’m not sure why Randall thinks the crater record is more reliable. 

The upshot of all this is that it may not be such a good idea to set aside work on extinction periodicity (even if only temporarily) in order to focus exclusively on the crater record. Even if we could show that the solar system's motion through the dark matter disk explains the periodicity in the crater record, we'd still be a long way from showing that it explains extinction periodicity. 

However, these critical questions should not diminish the coolness of a book that brings together the unobservables of paleontology with the unobservables of physics in such a fascinating way. And Randall’s story about the agency of the dark matter disk certainly has some potential to shed light on extinction.

Nemesis, perhaps, has struck back.


[1] Alvarez, LW, Alvarez, W, Asaro, F, and Michel, HV (1980). "Extraterrestrial cause for the Cretaceous–Tertiary extinction". Science 208 (4448): 1095–1108.

[2] Raup, D.M. and J.J. Sepkoski, Jr., “Periodicity of Extinctions in the Geologic Past,” PNAS 81(1984): 801-805.

[3] Davis, M., Hut, P., and Muller, R.A., “Extinction of Species by Periodic Comet Showers,” Nature 308(1984): 715-717.