An ugly hypothesis shot down by a beautiful fact

Figure 1. Radium Dial Painters

The simple truth that a single solid counter-example destroys any scientific hypothesis has been phrased many ways, none more memorably than Thomas Huxley in 1870 talking about how Pasteur took down Buffon and Needham's theory of spontaneous generation with a single experiment.

But the great tragedy of Science --- the slaying of a beautiful hypothesis by an ugly fact --- which is constantly being enacted under the eyes of philosophers, was played, almost immediately, for the benefit of Buffon and Needham.

Pasteur himself put it more prosaicly, ``Never will the doctrine of spontaneous generation recover from the mortal blow of this simple experiment."

The Linear No Threshold (LNT) hypothesis of radiation harm is the theory upon which our radiation protection regulation is based. In its pure form, LNT could be called beautiful in its simplicity. LNT is the theory that harm is strictly additive in the dose, the joules of radioactive energy deposited in a kilogram of tissue. One of the guys who pushed this idea was Harold Gray. We use his name as a shorthand for joules absorbed per kg tissue. Under LNT, we don't have to know anything about how slowly or quickly the dose was received. The only thing that counts is total dose in grays. This requires that the harm be linear in the total dose. According to BEIR VII, the bible of radiation protection, solid cancer incidence is 0.1455 *d for men and 0.2215 *d for women where d is the total dose in grays.\cite{beir-2006}[Table 12.5A, page 279]. What could be simpler?

But then things start to get a little jumbled. If the dose profile is "low dose", BEIR VII tells us to divide this incidence by a Dose and Dose Rate Effective Factor (DDREF) of 1.5. This leads to Table 1.

Strangely, we are not told at what Dose and Dose Rate we should stop applying the DDREF factor. Is there an abrupt jump at that point? It seems inescapable; but our LNT mentors are silent on the issue. In practice, LNT is incomplete. Reading between the lines a bit, it seems we should apply the DDREF in most situations not involving an atomic bomb.

To make matters still worse, even if we confine our attention to a single, quick dose, this simple model only sort of worked for photons and electrons. The workaround for other radioactive particles was to multiply grays by an RBE (Relative Biological Effectiveness) factor which factor depended on the type of particle and its energy. In other words, what we should really be worried about is the dose in sieverts which is RBE * grays.¹ LNT is the theory that harm is linear in the dose in sieverts, a complex function of absorbed energy, particle type, and particle energy.

This is no small matter. While the RBE for photons and electrons is 1.0 by fiat, the recommended RBE for alpha particles is 20. For neutrons, the RBE ranges from 1 to 20 in an odd manner depending on the neutron's energy. And this is just the top of the iceberg. The deeper you go the messier LNT gets. LNT no longer looks beautiful in its simplicity. It looks more like an ad hoc kluge. But the real problem with LNT is not its ugliness. The real problem is it's flat wrong. Like Pasteur, we need only one experiment to demonstrate this.

Between 1915 and 1950, numerals on luminous watch dials were hand painted using radium paint for the most part by young women. Prior to the late 1920's, the ladies used their tongues to form the tip of the brush into a point, sipping radium into their bodies. Chemically radium is similar to calcium and accumulates in the bones, where it has a 40 year biological half-life. The total skeletal doses varied by over a factor of 1000. But the maximum cumulative dose was an incredible 280 Gy.\cite{henriksen-2013}[p 276]

The Argonne National Lab did an extensive study of the results. 64 bone cancers and 32 head carcinomas were diagnosed. Reliable dose measurements were available for 2,383 women. All the 64 bone cancers occurred in the 264 women with a bone dose of more than 10 Gy.\cite{rowland-1994}[page 107] No bone cancers were found in the 2,110 women with less than 10 Gy dose.

Figure 2. Dial Painter Bone Cancer. LNT versus reality. On a log scale, LNT's straight line gets bent into a curve.

Figure 2 compares LNT's prediction and the actual dial painters bone cancer. Radium and its daughters are primarily alpha particle emitters. The top axis converts grays to sieverts, using an RBE of 16. Qualitatively, LNT looks nothing like the actual response. Quantitatively, even on a log scale, LNT is barely in the picture. At a total effective dose of 7 sieverts, LNT predicts every dial painter should have bone cancer. In fact, no cancers were observed in the 2,110 women who received up to 160 sieverts. If we asked the computer what are the ratios of the LNT cancer incidence to actual incidence, the answers would range from 168 to NaN (I don't have a number that large). In the rich history of bad predictions, this has to rank in the top ten.

64 preventable bone cancers is a terrible tragedy. But the ability of these women to handle massive amounts of radiation is a testament to our bodies' ability to repair radiation damage. That's a beautiful thing. I submit this is a case of an ugly theory being shot down by a beautiful fact. Regardless of the aesthetics, as soon as the dial painter data emerged, LNT, like spontaneous generation, should have been tossed immediately on the scrap heap of ideas that simply don't work. But that's not what happened.

LNT shot down but not killed

We know why LNT does not work. The fundamental reason that LNT performs so abysmally for the dial painters is it denies the ladies' ability to repair radiation damage to their DNA. DNA repair takes time. But for LNT the time dimension is irrelevant. LNT claims whether these women received their dose in one day or spread over 15 years makes no difference. LNT squashes decade long exposures into a single day.

This is Flat Earther level nonsense. Search ``DNA repair" on google scholar and you will get more than three million hits. DNA repair has been studied in mind boggling detail. We know an enormous amount about how DNA is repaired and how long it takes. LNTers simply refuse to accept any of this. This raises the question: why?

Almost all LNTers fall into one of two groups.

1) Anti-nukes. These people sincerely believe that nuclear power is evil. Their reasons range from a selfish, misanthropic fear of a prosperity driven population explosion to worry that widespread use of nuclear power will increase the likelihood of nuclear war. They know that neither concern resonates with the public. But fear of radiation is easily stoked, and LNT is excellent fuel for that purpose.

2)The LNT dependents. These are people whose livelihood depends on people being scared of radiation. This group comprises not just the radiation protection establishment, including the regulatory bureaucracies; but also the multi-billion dollar radiation clean up industry, the massive national labs researching solutions to all the LNT-inspired dangers associated with radiation, and the government agencies charged with doling out taxpayer dollars to pay for those solutions. Most importantly, it includes the industry incumbents. These people have spent hundreds of millions of dollars climbing over an LNT based barrier to entry. For these folks, the disappearance of that barrier would be a disaster. In short, practically all the existing nuclear power complex is dependent on LNT.

For both groups, accepting LNT's demise is not an option. But when it comes to defending LNT, the second group is far more important and effective. It contains almost all the "experts". The motives of the anti-nukes are obvious. Their claims automatically trigger scrutiny. But when an industry agrees with its opponents, case closed. LNT has no effective critics and survives, a triumph of self-preservation over Huxley's well deserved tragedy.

1. Rolf Sievert is one of the men responsible for the RBE fudge factor.