The Skeptics Society & Skeptic magazine

The great tragedy of science—the slaying of a beautiful hypothesis by an ugly fact.

—Thomas Henry Huxley

In recent weeks, physicist Lisa Randall has been promoting her new book, Dark Matter and the Dinosaurs. She even spoke about her latest work for the recent meeting of the Skeptic Society Science Salon on November 22. Naturally, any book which talks about such sexy topics as astronomy and dinosaurs is guaranteed to get lots of fawning press coverage, with little or no scrutiny from the scientific community. Nearly all the coverage and reviews of the book I have seen are either by science journalists without the appropriate background, or by astronomers and physicists. They were a bit skeptical about whether there was any strong evidence for her idea that waves of dark matter contributed to mass extinctions on earth, but could not rule it out. To her credit Randall made clear that she is proposing a hypothesis to be further tested, not a fully complete theory for which she is confident is correct. Like the good scientist that she is, Randall emphasizes that she could be wrong.

This is  not a review of the entire book, which is generally well written, and explains the complicated physics of dark matter and other topics is a clear and lively fashion. Unfortunately, the urge to tie her topic to sexy ideas like periodic extinctions and impact model for the end of the dinosaurs got the most press coverage, even if the book itself is more cautious about these topics. Still, it was a mistake to invoke the outdated and debunked ideas like the notion of periodic impacts causing extinctions, which seriously detracts from the credibility of an otherwise solid piece of science writing.

So far as I can tell, no one who actually knows much about geology or paleontology has reviewed or commented on her controversial idea. This is surprising, because there is now more than 35 years of research on the causes of mass extinctions in the geological past, and much of it directly contradicts her model. For one thing, her assertion that the impact at the end of the Cretaceous is the primary cause of the extinction of dinosaurs has been discredited in recent years. At the last two meetings of the Geological Society of America (Vancouver in 2014 and Baltimore in 2015), where over 6000 geologists and paleontologists meet to argue about topics like this, the consensus has now swung to the idea that the massive Deccan eruptions in India and Pakistan were far more important to the end-Cretaceous extinctions. Randall gives a brief discussion of the Deccan eruptions (pp. 202-203), but does not accurately reflect the consensus view of the geological community about their great importance to the end-Cretaceous extinctions. Based on all the recent literature in geology, and the talks given at recent geology meetings, the impact of an extraterrestrial object (whether asteroid, comet, or dark matter) has been considerably reduced in importance. Yet much of the general public is unaware of this changed conclusion in geology. The popular media (and even the science media) still propagate the simplistic notion of the rock from space doing all the damage, without mentioning the other causes that are even better documented, or the complexity of the pattern of extinction and survival.

Artist’s conception by NASA artist Don Davis of the impacting asteroid smashing into earth at the end of the Cretaceous. (Via Wikimedia Commons).

Even more controversial is her assertion in both the book and the interviews that the dark matter model might be the cause for periodic extinctions in the fossil record. Hearing this statement is as jarring to a geologist or paleontologist as going through a time warp. The periodic extinction model was first proposed in 1984, but has been completely debunked since 1990, and almost no geologist or paleontologist takes it seriously any more (except for the maverick Mike Rampino, who is quoted extensively in the interviews and cited in the book, even though no one else in geology follows him).

The initial idea of periodic extinctions was first published by the late David Raup and Jack Sepkoski in 1984 (Raup and Sepkoski, 1984, 1986; Raup, 1986, 1991). While their paper was still circulating as a preprint, a number of astronomers jumped the gun before the idea had even been published or received proper scientific assessment. These astronomers were quite imaginative in proposing causes for this alleged periodicity. They postulated periodic comet showers (Davis et al., 1984), the oscillation of the solar system through the galactic plane (Rampino and Stothers, 1984; Schwartz and James, 1984), an unknown Planet X (Whitmire and Jackson, 1985), and even an undetected companion star to the sun named Nemesis (Whitmire and Jackson, 1984). Loper and McCartney (1986) and Loper et al. (1988) suggested that there was a 26-million-year periodicity in mantle overturn within the earth, triggering pulses of volcanism and global climate change that then caused extinctions.

Unfortunately for the pro-impact stampede, several ugly little facts killed their beautiful hypotheses. No evidence for Nemesis or Planet X has ever been found. Randall freely admits this on p. 260. So why does she even mention “Nemesis” again, 30 years after it was debunked and vanished from the scientific literature? Nor has any evidence tied comet showers or the motion through the galactic plane to mass extinctions (Shoemaker and Wolfe, 1986; Tremaine, 1986; Sepkoski, 1989). Randall spends the entire Chapter 15 reviewing this topic, and confesses there is no evidence for it.    Similarly, the mantle periodicity model has been discredited. In fact, the very existence of the 26-million-year extinction cycle has been challenged on statistical grounds (Kitchell and Pena, 1984; Kitchell and Easterbrook, 1986; Hoffman and Ghiold, 1986; Harper, 1987; Stigler and Wagner, 1987; Noma and Glass, 1987; Quinn, 1987). Randall (p. 244) admits that the statistical support for the periodicity model is very poor, but why then does it get so much coverage?

Cladistic taxonomists have criticized Sepkoski’s database because it is full of paraphyletic or monotypic taxa that are not real monophyletic groups, as well as bad taxonomy and misidentifications. When echinoid specialist Andrew Smith and paleoichthyologist Colin Patterson examined the database for their taxa of specialization (echinoderms and fishes) and eliminated the mistakes and non-monophyletic groups, the periodicity disappeared (Patterson and Smith, 1987; Smith and Patterson, 1988).

Another problem with the data is the way they are compiled. Since the quality of the data is highly variable, Sepkoski lumped all the data by stages of 3 to 5 million years in duration. This means that all extinctions that happened at different times within a given stage are treated as if they occurred exactly at the end of the stage, even if they were evenly spaced through the duration of the stage. Such a method artificially bunches all the extinctions at stage boundaries and makes a gradual extinction pattern appear catastrophic. Randall (p. 173) discusses the Signor-Lipps effect and how it might make an abrupt extinction appear more gradual, but does not seem to recognize this serious compilation flaw in Sepkoski’s data base from which the entire periodic extinction model arose.

The dating is not very reliable either. The time scales have changed so much in recent years that the 26-million-year prediction can succeed or fail depending upon which time scale is used. For example, Raup and Sepkoski (1984, 1986) predicted a late Eocene extinction peak at 39 Ma, and at the time, the age of the Eocene/Oligocene boundary was disputed, ranging from 36.5 to 32 Ma. Even with time scales in use in 1984, it appeared that their prediction was off. Today, we place the middle/late Eocene extinction at 37.2 Ma, the Eocene/Oligocene boundary (not much of an extinction) at 33.9 Ma, and the earliest Oligocene extinction at 33.0 Ma, so none of these match Raup and Sepkoski’s (1984, 1986) prediction. Randall (p. 233) mentions the Eocene impacts briefly, but does not seem to be aware of the literature that shows NO extinction at the time of this impact (33.5 Ma, in the middle of the late Eocene). This demonstrates that even very large impacts (the Chesapeake and Popigai impacts in the late Eocene were only slightly smaller than Chicxulub impact that came at the end of the Cretaceous) can cause NO extinctions.

The biggest problem with the periodic extinction model, however, is the fact that there is no common cause for the major mass extinctions, which would be required if the same triggering event occurred with a regular pattern. Only the end-Cretaceous extinction is associated with an impact, but this is true of no others (despite all sorts of false alarms over the years). Gigantic mantle eruptions are associated with the end-Permian, end-Triassic, and end-Cretaceous extinctions, but with no others. The late Ordovician and late Devonian extinctions have no clear indication of impact or volcanism, but seem to be due to global cooling. The middle and late Eocene extinctions at 37 Ma and 33 Ma were also apparently due to global cooling, but (as already mentioned) the late Eocene impacts happened BETWEEN the extinction horizons and caused NO extinctions. Randall (p. 179-185) even reviews the evidence for most of these mass extinctions, but there is no discussion how this lack of common cause invalidates the entire periodicity model. Only on p. 243 does she seem to indicate that extraterrestrial impacts are not so important–but then why spend all the time and pages talking about periodic extinctions caused by extraterrestrial events as if they had any validity?

Finally, there is a serious question whether many of the extinction “peaks” are real. The middle Miocene “extinction peak” at 13 million years is based on a few species of molluscs and does not show up in the excellent record of land mammals (Webb, 1977; Barry, 1995; Heissig, 1979). Randall (p. 233) mentions the impact found at this time (Ries Crater in Germany), but not the fact that it caused no mass extinction, even in the land mammals from the immediate vicinity (Heissig, 1979). The early Jurassic peak was barely above background noise levels, and Sepkoski (1989) abandoned the mid-Jurassic extinction peak. Some extinction peaks (the late Triassic, the mid-Jurassic, the early Cretaceous, and Pliocene) fall well outside the predicted time interval (Sepkoski, 1989). If only half of the “peaks” appear to be real and occur on schedule, and there are long gaps with no extinction at the predicted 26-million-year interval, what does this imply about the “periodicity”?

Stanley (1990) suggested a much simpler explanation for this apparent periodicity. Major mass extinctions tend to kill the highly specialized taxa, leaving only extinction-resistant generalists known as “survivor” taxa. In the aftermath, it takes many millions of years for diversity to recover and for a wide variety of extinction-prone specialists to evolve and fill the vacant ecological niches. If some extreme event occurred soon after a major mass extinction, there would be no significant extinctions, because the only organisms alive would be the extinction-resistant “survivor” taxa. Only after 10 to 20 million years does diversity return with specialized taxa that would be vulnerable to a major environmental perturbation. The 26-million-year “periodicity” may simply be a reflection of the time it takes for a fauna to recover before it can feel the effects of the next climate change or major eruption or other event. This would also explain why the “cycles” are not precisely 26 million years, but vary in duration. An astronomically-caused cycle would be much more regular.

Although the periodicity model was very popular and influential in the late 1980s and 1990s, today it is regarded as a historical curiosity that has not withstood the test of further scrutiny by scientists. Randall (p. 233) seems to be aware of this problem, but never fully comes to terms with it. Meanwhile, she succumbs to the temptation to talk about dinosaurs and periodic extinctions which make the topic more glamorous. It is disappointing that such an otherwise well-written book spends so much time on an idea debunked more than 20 years ago.