Ed Lewis' Dilemma

If LNT is biological nonsense, how did it ever get accepted? The Hamlet in that tragedy is Ed Lewis.

The Rockefeller Foundation and the Genetic Scare

On August 20, 1945, Ernest O. Lawrence wrote to the Rockefeller Foundation (RF) thanking them for their critically important help in developing the atomic bomb. Lawrence said ``that if it had not been for the RF, there would have been no atomic bomb".\cite{lawrence-1945} Lawrence was probably right. In their support of theoretical physics in the 1930's, the RF had funded just about all the Manhattan Project greats. Much worse, they had single handedly funded Lawrence's cyclotron program, which turned out to be crucial in developing the bomb.

Foundation President Raymond Fosdick, a lawyer who was involved in setting up the League of Nations, was not happy. On the 29th, he wrote to Warren Weaver, the RF's Director for Natural Sciences, saying ``his conscience was deeply troubled".\cite{fosdick-1945} Fosdick and Weaver decided to make amends and do whatever they could to control nuclear weapons, starting with ending weapons testing. Here's what Fosdick told the Rockefeller Trustees later that fall.

Whether the release of atomic energy in the long run will result in good or evil for the race, no one can now say; but whatever the consequences, the Foundation and its related boards cannot escape their share of the responsibility, indirect as it may be. The atomic bomb is the result of influences which, for the most part unintentionally and unwittingly, we helped to set in motion.

...

The towering question which faces the world now is whether the new energies can be controlled. It is, I know, the hope of all of us that the Foundation may be able to make some contribution, however slight, to this end.\cite{rf-1956}

Fosdick and Weaver got right to work. Their problem was that with a few local but dramatic exceptions, the dose rates resulting from weapons testing were well below background. The solution was to argue that mortality was linear in dose however small and cumulative over both time and population. The solution was LNT. If they could get LNT accepted, then they could aggregate these tiny dose rates over long periods and hemispherical populations, and claim that bomb testing was invisibly killing millions.

They had the weapon they needed close at hand. The Associate Director of the Natural Science Division which funded genetics research was Frank Hanson.¹ Hansen had worked with Herman Muller and in early 1945 had gotten Muller a job at Indiana University by effectively paying for his $6000 per year salary.\cite{carlson-1981}[p 286-287]² In 1946, Weaver approved a massive increase in the Foundation's funding of Muller's work at Indiana. By 1956, the Foundation has spent $3,781,000 at Indiana, in support of Muller's work.\cite{wynchank-2011}

In 1927, Muller had shown that X-rays could produce mutations in Drosophila fruit flies. In 1930, Muller had claimed that the mutation frequency ``is exactly proportional to the energy of the dosage observed" despite the fact that his own data did not support linearity, and in 1927 and 1928 papers he discussed the implications of the non-linear response.\cite{calabrese-2009}[page 206] This claim was based on his theory that a single change in a gene, which Muller called a `point mutation' or a `hit', caused the big changes that Muller observed in his flies. We now know that the large doses, 2750 mSv or more in periods of an hour or less, that Muller was basing his judgement on induced massive gene deletions in the flies.\cite{fossett-1994}

Also in 1946, despite his rocky academic career, Muller was awarded the Nobel prize. Five weeks before he received his award, Muller received a manuscript from Ernst Caspari, a fruit fly researcher he knew well. Caspari had been given the job of confirming that Muller's linear, dose rate independent rule extended down to dose rates 2500 times lower than had been tested at the time. He irradiated a group of flies at 25 mSv/day for 21 days. He meticulously maintained a control group under exactly the same conditions, except for the radiation. The test was female sterility. To Caspari's consternation, there was no statistical difference between the irradiated females and the non-irradiated, Table 1. This should have been a bombshell.

Table 1. Caspari Table 2. 52.5 r is 525 mSv. Gamma-rays are photons.\cite{caspari-1948}}

Caspari worked in Curt Stern's lab. Muller wrote to his buddy Stern admitting he could find no problem with Caspari's work, only asking that it be repeated.³ Yet a few days later in his Nobel acceptance speech, Muller claimed:

They leave, we believe, no escape from the conclusion that there is no threshold dose, and that the individual mutations result from individual hits, producing genetic effects in their immediate neighborhood.

The RF made sure Muller received plenty of publicity, funding speaking trips all over the world.

In 1954, the Foundation contracted with the National Academy of Sciences (NAS) to perform a review of the biological effects of radiation. Under this contract, the NAS set up the Biological Effects of Atomic Radiation Genetics Panel (BEAR). Warren Weaver was put in charge of the Genetics committee. Weaver stacked the committee with laboratory biologists, most of whose work was done on fruit flies, and much of that work was funded by the Foundation. Radiotherapists who had worked with humans need not apply.

Muller was the prominent member of the Genetics committee, arguing strongly for linearity which would be an abrupt departure from the prevailing position that there was a tolerance dose rate below which there would be no detectable harm, a position which was consistent with Caspari's results, which the Panel simply ignored. The Panel held that genetic damage was unrepairable and therefore the damage just kept adding up, and harm and damage were equivalent. Therefore, harm depended only on total dose. Dose rate was irrelevant.⁴ If this is the case, harm must be linear in dose.

The key decision by BEAR to accept LNT was made at a February 6, 1956 meeting with little or no debate.\cite{calabrese-2019}[page 13] At the meeting, Weaver made sure everybody understood what was at stake. He told the group that he would ``try to get a very substantial amount of free support for genetics, if at the end of this thing, we have a case for it. I am not talking about a few thousand dollars, gentlemen. I am talking about a substantial amount of flexible and free support to geneticists". The Foundation was quite prepared to use the geneticists' cupidity to induce scientific misconduct, if that's what it took to stop nuclear weapons testing.

Arthur Sulzberger, publisher of the New York Times, was also a member of the RF Board of Trustees. The New York Times immediately ran a front page story with the headline ``SCIENTISTS TERM RADIATION A PERIL TO THE FUTURE OF MAN". The paper carried a series of articles amplifying and at times exaggerating the Panel's findings. The Foundation's plan was going well.

But there was a problem. Starting in 1946, the government had funded the same National Academy to do a study of birth defects in children born to atom bomb survivors. The leader of this study was James Neel. Over 10 years, 70,000 pregnancies were studied. In 1956, the NAS published the results.\cite{nas-1956} There was no evidence of any harm to children conceived after the bombs were dropped.⁵

The Genetics committee was aware of the Neel study which issued periodic reports on its progress. But they chose to ignore it, preferring censored, fruit fly data over human data. As Muller put it, ``We should beware of reliance on illusionary conclusions from human data, such as the Hiroshima-Nagasaki data, especially when they seem to be negative". But after publishing the full report, Neel took his data to Europe, where he found a much more receptive audience. British scientists generally accepted the Neel study and it became part of a major WHO report, despite aggressive threats from Muller.⁶ The Genetics committee defense by dismissal was not working.

Fortunately, a better solution soon appeared. In May, 1957, a fruit fly biologist, E. B. Lewis, who had studied under a Muller protege published a paper in Science, claiming a relationship between radiation dose and leukemia.\cite{lewis-1957} And the relationship was linear and cumulative, just like Muller's fruit fly model. We will take a look at Lewis's methods shortly.

Lewis's paper created an avalanche of favorable publicity, including a gushing editorial by Science's editor-in-chief, Graham DuShane. DuShane was quite clear about why he was so pleased with the paper: ``Thanks to Lewis, it is now possible to calculate --- within narrow limits --- how many deaths from leukemia will result in any population from any increase in fallout or other source of radiation."

The National Academy switched its focus to cancer. The Biological Effects of Atomic Radiation Genetics Panel label was quietly dropped and replaced with the Biological Effects of Ionizing Radiation (BEIR). As Muller predicted, the Rockefeller Foundation stopped funding fruit fly research. The theory of genetic harm to humans from radiation lived on mainly in low budget horror flicks. But the pre-DNA genetic hypothesis that radiation damage was unrepairable and therefore all that counts is the total dose survived.

Lewis' Lie

About the same time, Science published Lewis' bombshell paper, the Atomic Bomb Casualty Commission (ABCC) published their own analysis of essentially the same data. The ABCC (later the Radiation Effects Research Foundation) was set up by the AEC to study the effects of radiation on the bomb survivors. Table 2 reproduces Table VII from the UNSCEAR 1958 report except that I have converted the dose in rem to mSv. The lettered footnotes are in the original.

This table makes a number of points.

1) The Zone A and B exposures increased leukemia incidence by a factor or 40 and 14 respectively. Leukemia is a rare disease. As of 1957, 12 cases has been diagnosed among the 66,000 people in Zones D and E. Pretty clearly, almost all the leukemias in Zones A and B were caused by radiation.

2) The response was highly non-linear in dose. If we take the average zone doses at face value and use the Zone E rate as background, the excess incidence per mRem (0.01 mSv) is 0.091 for Zone A, 0.069 for Zone B, 0.025 for Zone C and -0.9 for Zone D. Since the grouping and averaging by zone washes out a lot of the non-linearity, we can be sure the actual numbers were even more non-linear.

3) The 32,692 people in Zone D had a lower leukemia rate than the 32,963 people in Zone E. The zone with the higher average dose had less disease than the zone with the lower. 66,000 people is a large sample.

So how did UNSCEAR interpret this? Again I've converted rem to mSv in the quotes.

In zones A (13,000 mSv), B (5000 mSv) and C (500 mSv), the values of P_L were calculated to be 0.09, 0.07 and 0.07 [sic] times 10^-6 respectively. This finding was taken to support the suggestion that the extra leukemia incidence is directly proportional to radiation dose, and conversely to argue against the existence of a threshold for leukemia induction.\cite{unscear-1958}[para 31, page 165]

P_L is the extra probability of leukemia occurring per dose since exposure, the last column in Table VII. So 0.091, 0.069 and 0.025 (without the typo) are equal? What about the Zone D numbers?

Contrary to previous findings, the present findings indicate that P_L decreases markedly as the dose falls, that therefore leukemia incidence is not a linear function of dose, and that a threshold for leukemia induction might occur. In fact according to Table VII, a dose of 20 mSv is associated with a decreased leukemia rate. It is to be emphasized again, however, that estimates of dose employed in the present and previous analyses are much too uncertain to permit drawing conclusions relative to the vital points in question. The calculations are made only to illustrate how variable the results may be when inadequate data are utilized.\cite{unscear-1958}[para 33, page 165]

In other words, the uncertainties are such that the numbers can be ignored; but they support LNT even when it looks like they don't. This is Wonderland stuff.

But wait a minute. How did Professor Lewis, working from essentially the same data come up with a linear relationship between dose and leukemia? At the heart of Lewis's argument is his Table 2 reproduced here as Table 3.

There are some differences in the population. The UNSCEAR table refers only to Hiroshima survivors. Lewis combined Hiroshima and Nagasaki. But semi-quantitatively the results are somewhat similar, with one very striking difference. Lewis has lumped Zone E and D together while UNSCEAR did not. In so doing, Lewis hid the glaring non-linearity in the data.

Lewis knew exactly what he was doing. To defend his decision to include Zone D into his control group, he feels compelled to say ``the average dose is under 5 rem [50 mSv] and is thus so low that zone D can be treated as if it were a `control' zone.''\cite{lewis-1957}[page 125] But if there is no difference between 50 mSv acute and zero, the relationship cannot be linear. The widely acclaimed and enormously influential Lewis paper was not only deceitful, it was inconsistent. The UNSCEAR explanation might be gibberish; but at least they did not hide the data.

Why did Lewis lie?

Lewis was a very smart guy. He later won a Nobel prize. While Muller comes across as a jerk, by all accounts Ed Lewis was an excellent human being. For 45 years, he was a popular and much loved faculty member at Caltech. And his lie is so transparent! Why did nobody point it out? The AEC had funded both the Caspari study and the ABCC study. They knew Muller's claims were rubbish. They knew the ABCC leukemia data was highly non-linear, and must have known what Lewis had done. Hundreds of scientists could have pointed out the glaring inconsistency. But as far as I know none did. Moreover, fractionation, dividing a therapeutic dose into fractions, delivered a day or so apart to allow healthy cells to recover, was universal medical practice. If LNT is valid, fractionation makes no sense. Even Lauriston Taylor, a towering figure, who called LNT "a deeply immoral use of our scientific heritage" did not speak out until 1980. He was about 25 years too late. Were all these people grasping creeps?

Of course not. They were petrified of the bomb. If LNT could end bomb testing, then I will have to abandon scientific integrity, just this once. Look at Taylor's strange wording. You don't normally call a model, a "use". He knew LNT had been accepted not because it was correct, but because it was a tool, a tool for controlling the bomb. But by 1980, the auto-genocidal downside of the well-intentioned fraud had become apparent. I'm confident Taylor included himself in that "immoral use".

Were all the people who broke all the rules of scientific integrity to push LNT, or at least remain silent when they knew better, wrong to do so? I don't know. Would I have done differently? I don't know. But I do know that the reason LNT was accepted with essentially zero pushback despite the fact that it is biological nonsense was fear of the bomb.

1. The Foundation had a long history of funding genetic research, motivated at least in part by their interest in eugenics. As late as 1952, Fosdick would say that the foundation’s investments in natural sciences were still guided by the questions it had begun asking in the 1930s, like whether it was possible to “develop so sound and extensive a genetics that we can hope to breed in the future superior men.”\cite{fosdick-1952}[p 166]

2. Why Indiana? Muller was an abrasive character who turned people off wherever he went. He bounced from Rice, to Columbia, to Texas University where he dabbled in Communism and eugenics. After a suicide attempt in 1932, he moved to Germany and then the USSR. He spent 5 years in Russia; but, because he was on the wrong side of the Lysenko argument that genes were malleable by the environment, was lucky to get out alive. He moved to Edinburgh and then Amherst during World War II. At the end of the war, Amherst informed him that he was being fired. Frantic letters to colleagues produced nothing.

The Rockefeller Foundation money was his salvation. But finding a place to spend that money proved a problem. The combination of Muller's personality and his Communist past made him unwelcome just about everywhere. Indiana was an exception.

Ironically Muller ended up a bit of a Cold War hawk. His time in Russia had soured him on Communism and taught him to be suspicious of some of the more extreme disarmament proposals.

3. When Caspari finally published his results in 1948, he treated them as an anomaly, something to be studied further.\cite{caspari-1948} Caspari did comment that, if his results were proved correct, they would be consistent with the sigmoid response seen in the killing of bacteria and Drosophila eggs by radiation. Caspari realized he had not proven there was a threshold in the strict sense. But his results most emphatically contradicted Muller's linearity claim.

4. Muller and his colleagues were the last generation of classical biologists. They lived in Mendelian world of on-off genes. In such a framework, there is no mechanism for repair. But that model was already being superseded by a far richer, far more descriptive understanding based on the discovery of DNA in 1953.

5. There has been a series of follow up studies extending into the 1990's.\cite{nas-1991} They have confirmed and strengthened the original results.

6. The acrimonious correspondence shows that Muller was much more worried about funding than pushing LNT.\cite{calabrese-2020} Neel was challenging the whole idea of using fruit fly data to predict human response. If Neel was right, Muller's funding would dry up.