The scientist would then insert a device to extract sediment samples at different stratigraphic levels above, below, and at the boundary. These would be taken back to the lab, dated and analyzed. The analysis would require separating and counting micro- or nano-sized particles using a high-powered microscope, all the while avoiding contamination and misidentifications. To distinguish extraterrestrial microspherules from terrestrial requires special instruments like the scanning electron microscope and X-ray spectrometer. Moreover, unless the original methods that led to the discovery of the event markers are followed precisely, the would-be replicators could well get a different result. Remember that the microspherules are just that: so small they cannot be seen with the naked eye.
GUBBIO
No one expects a geologist or archeologist to replicate his or her own work by returning to some distant outcrop or ancient archeological site and sampling it all over again from scratch. Then no field scientist would get anything new done, nothing would ever be final. In the case of the Alvarez Theory, no one suggested that Walter Alvarez should return to Gubbio, Italy, where he and his father had discovered the critical iridium spike, sample the KT boundary there again, return the samples to the lab at UC Berkeley, rerun the neutron activation tests for iridium, and so on. Instead others applied a more useful kind of reproducibility test: to see if the iridium anomaly also occurred at known impact craters and at other KTB sites. It did, thus indirectly confirming the reproducibility of the iridium spike at Gubbio without anyone having to repeat the exact same measurements. This approach had the obvious advantage of showing iridium to be a worldwide event marker of extraterrestrial events.
Here is a re-plot of the data from the 1980 article by the Alvarezes and colleagues. (The yellow line is a drawn-by-hand fit to the data points.) Not even the most vocal opponents of the Alvarez Theory ever suggested that these findings were false or felt they needed to replicate them. Rather they argued that the iridium spike was due not to impact, but to volcanism. That was not an unreasonable argument, but the evidence failed to support it.
FIGURE 4:
The iridium peak at Gubbio, the clue that led to the Alvarez Theory.
SELF-REPLICATION
Instead of repeating their laboratory tests multiple times on samples from a single YD site, FEA used the same method at ten different sites — as though the Alvarezes had themselves sampled nine other KTB sites in addition to Gubbio, instead of leaving that to others. By finding similar event marker peaks at each, FEA demonstrated that they could reproduce their own results and that the peaks are not a one-site fluke, but are common to sites in different geological settings separated by thousands of miles.
As we have noted before, the FEA paper met several criteria of good science. It:
• Appeared in one of the most selective journals, having received the approval of its peer-reviewers and editors.
• Replicated its own procedures and findings.
• Appealed to a known and accepted process: meteorite impact.
• Was written by more than two dozen respected and well-published experts in a variety of fields.
• Described its methodology in detail both in the article itself and in online supplemental materials.
• Did not overreach in its claims, saying, “Our primary aim is to present evidence supporting the YD impact event.”
UNABLE TO REPRODUCE
SEA set out to determine whether “the supposed markers of extraterrestrial impact peak only at the onset of the YD in sedimentary contexts across a broad geographic area. In other words, are the Firestone et al. results replicable?” They studied the “magnetic minerals and microspherules from seven sites of similar age [to the YDB], including two examined by Firestone et al.”
But they were “unable to reproduce any results of the FEA study,” even at the two sites in common: Blackwater Draw and Topping, S.C. This even though at Blackwater Draw, they had “collected within a few centimeters of the sections sampled previously [by Firestone et al.]” and even “us[ed] methods from the original FEA study.”
The charts below present the results of each study for Blackwater Draw. The scales differ and SEA present their results as ranges rather than data points, but it is plain to see that where FEA found peaks in the abundance of magnetic grains and spherules at the YDB, SEA found none. As shown, SEA did find microspherules at Blackwater Draw, but only above the boundary.
FIGURE 5:
Firestone et al. find abundance peaks in event markers at Blackwater Draw.
FIGURE 6:
Surovell et al. find no abundance peaks in magnetic grains and spherules at the YDB (12.9) at Blackwater Draw.
Here lies a conundrum. After all, the microspherules are physical objects, albeit minuscule, and the peaks reported at Blackwater Draw either exist or they do not. One group of authors found them and the other did not. Whose result should we accept? We need to think this through carefully not only to judge whether the initial accusations of irreproducibility against the YDIH were fair and correct, but also because this example provides a case study of the issue of reproducibility in science.
Short of accusing one side or the other of dishonesty, there appear to be only two ways to explain these discrepant results: (1) the methods of FEA repeatedly created false peaks where none exist, or (2) the methods of SEA caused them to miss or to destroy peaks that do exist. The two have an entirely different probability of occurrence: the first is hard to conceive, the second is a constant worry for those researching nano-sized objects. Nevertheless, opponents of the YDIH endorsed the first option. The scarlet letter of irreproducibility would follow the YDIH right up to the present day.
POTENTIAL ERRORS
It is easy to understand how errors could creep in to the analysis of the microspherules, due to:
1. Sampling that missed the YDB. For instance, the YDB layer at Blackwater Draw, as estimated from the base of the peaks shown in the chart, looks to be about 15 cm (6 inches) wide. Sampling at 20 cm intervals could cause one to miss it entirely. Or, since it is obscure at some sites, it might simply not have been recognized.
2. Collecting too few samples for statistical significance.
3. Using a magnet that failed to extract all the microspherules, or using different magnets of unequal power.
4. Losing microspherules in the washing and cleaning process.
5. Failing to recognize and count ET microspherules.
6. Incorrectly identifying impact-derived microspherules as terrestrial.
But — and this is crucially important to our story — such errors could not have manufactured peaks where none existed. Had the microspherules been randomly distributed across the YDB, mistakes and misidentifications by FEA could not have created peaks, but only added to the disarray. Mistakes destroy order, they do not create order out of randomness. On the other hand, had the peaks been real, errors would have blurred or destroyed them. Had the Alvarez team jumbled some of their iridium results, say, that would have destroyed the iridium peak. This is why I say these two kinds of errors have a quite different probability.
Moreover, if mistakes by the FEA authors had created a false peak in the microspherules at Blackwater Draw, then those same sorts of mistakes must have produced false peaks at each of the other nine sites they studied, revealing a fundamental and systemic flaw in their methodology. But no critic has ever identified what that flaw might have been. Moreover, as we will see, scores of scientists have used the FEA techniques and replicated their finding of microspherule peaks.
In Chapter 8, I will come back to the question of whether SEA did make mistakes that might have caused them to miss the peaks that FEA found. For now, I want to ask only whether it was logical and fair for the scientific community to allow SEA’s reported absence of evidence to trump the positive evidence reported by FEA and to label the latter’s results “irreproducible.”
In a concluding section, SEA seemed to acknowledge the possibility that they might
have erred:
Assuming an ET impact occurred, perhaps the lack of reproducibility indicates that the methods used for recovering the magnetic material are not appropriate for the task at hand. Recognition and identification of the spherules is especially difficult and somewhat subjective.
This was an honest assessment that one might have expected would lead SEA to temper their conclusions and declare the matter still open. Instead they wrote, “Replicability is fundamental to the scientific method and hypothesis testing; results that are not reproducible cannot be considered reliable or supportive of a hypothesis.” This is close to a tautology, but in the context of their article it creates the clear implication that it is the results of FEA, and not their own, that are irreproducible.
They conclude: “We find no support for the extraterrestrial impact hypothesis as proposed by Firestone et al.” This again is literally true, but the clear implication in context is that no such support, no such evidence, exists.
The choice of words clearly labels the YDIH evidence “irreproducible.” But SEA did not establish that. Rather they showed only that they could not reproduce one specific result: the microspherule peaks. How can anyone be sure that is was not SEA’s results that were irreproducible?
Suppose that SEA had published first, reporting that they found no peaks in event markers at the YDB. Then suppose that FEA subsequently reported finding those very peaks. The second-arriving positive evidence of the peaks would likely have overruled the earlier-reported absence of evidence and it would been SEA’s results that were labeled “irreproducible.” But surely the date of publication cannot be the criterion of reproducibility.
Opponents of the YDIH cited SEA’s allegation of irreproducibility over and over, without ever raising the questions that we have raised in this chapter. (See deadlyvoyager.net for examples.) The accusation dug the hole even deeper for the YDIH, leading opponents to go so far as to liken it to pseudoscience.
6
A CAUTIONARY TALE
After SEA published in 2009, several articles critical of the YDIH appeared, while others replicated FEA’s findings, reported new evidence, and rebutted the criticisms. For example, William Mahaney and colleagues found the black mat in the Venezuelan Andes, and attributed it to a cosmogenic origin. At the Clovis site at Sheridan Cave, Ohio, two scientists reported, “Lonsdaleite, nano-diamonds, magnetic microspherules, magnetic grains, and carbon spherules [that] were absent in layers above and below the Clovis assemblage.” As we saw in Chapter 4, a group reported a layer of nanodiamonds in the Greenland ice sheet.
Thus by mid-2011, the time was ripe for a balanced review article that addressed the allegations of irreproducibility, assessed the evidence for the YDIH, and suggested directions for further research. And indeed, in that year Earth Science Reviews published, “The Younger Dryas Impact Hypothesis: A Requiem.” It declared the hypothesis dead and good riddance:
Throughout the arc of [the YDIH], recognized and expected impact markers were not found, leading to proposed YD impactors and impact processes that were novel, self-contradictory, rapidly changing, and sometimes defying the laws of physics. The YD impact hypothesis provides a cautionary tale for researchers, the scientific community, the press, and the broader public.
If a title should say clearly what an article intends to show, this one succeeded. According to the Oxford English Dictionary, a Requiem is “A mass said or sung for the repose of the soul of a dead person.” It’s use in the title of a scientific article is tantamount to saying that the hypothesis in question has been shown to be false and that scientists would be foolish to pursue it.
But is a label like “Requiem” ever a proper term to apply to a scientific hypothesis? The history of science answers firmly: no. Consider these examples of premature burial:
• In 1944, at age 85, the influential and well-traveled geologist Bailey Willis published, “Continental Drift, ein Märchen (a Fairytale).” He urged that continental drift “be placed in the discard, since further discussion of it merely incumbers [sic] the literature and befogs the minds of fellow students.” Willis aimed to erase continental drift from the very curriculum of geology and from the minds of geologists.
• The eminent geologist Walter Bucher also urged geologists to forget continental drift: “The student must get along without assuming either that vast continental areas have foundered to form large areas of the ocean floor, or that continents and islands have drifted.”
• Bucher also said that, “All American cryptovolcanic structures [now known to be impact craters] represent special phases of the ascent of basic magmas into the central plateau region.”
• In The Great Dinosaur Extinction Controversy, Charles Officer wrote that “One of the things that did not happen at the KT boundary was an impact by a gigantic meteorite.” And, “The Alvarez hypothesis has collapsed. [It is] degenerate...not merely pathological science but dangerous to boot.”
• Three different authors provided a death certificate for Arrhenius’s CO2 theory, saying:
The carbon dioxide theory has failed utterly under searching criticism.
Arrhenius did not develop a geological theory but merely made an advance on Tyndall’s suggestion as a physicist, tho’ the “advance” proved unfortunate.
The theory was never widely accepted and was abandoned when it was found that all the long-wave radiation absorbed by CO2 is also absorbed by water vapour.
Theories can be born prematurely, but so can they be buried.
PATHOLOGICAL PSEUDOSCIENCE
The abstract of the Requiem is nearly 600 words long and includes the paragraph above. Anyone seeking to learn the state of the YDIH, having read only the abstract and finding the hypothesis so thoroughly disparaged, would have seen no reason to read further.
The article itself proceeds to a detailed review of the various kinds of evidence on the YDIH. A major section is on the authors’ research on the Santa Barbara Channel Islands, including Arlington Canyon on Santa Rosa Island, the YDB site that James Kennett and colleagues had written about in several articles. But in contrast to those articles, like SEA, the Requiem authors could find no microspherules, no evidence of an extraterrestrial event at Arlington Canyon or at the other Channel Islands sites they sampled. This led them to question the sampling methods of the proponents of the YDIH, but as we will see, others would argue that the Requiem authors had themselves failed to sample the YDB, which if true would render their results meaningless.
A section near the end of the Requiem has the same tone as the GSA Today article and needs to be quoted in full:
Scientific hypotheses are constantly being proposed, tested, confirmed, or cleanly rejected, but a small minority of these stray from this time-proven path. Many scientists are unaware of the surprising number of hypotheses that have gone badly astray, often after widespread initial interest and support (Langmuir and Hall, 1989; Gratzer, 2000; Park, 2000). Characteristics of these wayward hypotheses include claims that are spectacular, data that are subjective or at the limit of precise measurement, and criticisms met with ad hoc excuses and/or shifts in the original claims. We suggest that much can be gained by stepping back and looking at the broader lessons for the earth sciences, impact science, archeology, and other affected fields (Italics added).
What are examples of these “wayward hypotheses” and the “broader lessons” they convey? The answer must lie in the three sources the authors cited.
Langmuir and Hall (1989) refers to an article in Physics Today titled Pathological Science. It is a transcript (by Hall) of a legendary talk given in 1953 by Irving Langmuir, winner of the 1932 Nobel Prize in Chemistry. For anyone interested in the history of science, Langmuir’s talk makes fascinating reading.
He identified six “symptoms of pathological science.” Two apply mostly to the laboratory sciences, but four pertain more generally:
• Claims of great accuracy.
• Fantastic theories contrary to experience.
> • Criticisms met by ad-hoc excuses thought up on the spur of the moment.
• The ratio of supporters to critics rises up to somewhere near 50% and then falls gradually to oblivion.
Langmuir provided a number of examples of science gone wrong, including the illusory N rays, extrasensory perception, and flying saucers. Cold fusion would later come to be seen as a modern example.
Some of these, Langmuir said,
Are cases where there is no dishonesty involved but where people are tricked into false results by a lack of understanding about what human beings can do to themselves in the way of being led astray by subjective effects, wishful thinking or threshold interactions.
Deadly Voyager Page 4
