In the previous post, I tried to make sense what Compton (in a 1959 interview with Pearl Buck) and Bethe (in 2000, reported here) said about the analyses of the possibility that the first nuclear tests might ignite the atmosphere. Independently, on the same day, John Horgan wrote an excellent blog post on the same topic with many more interesting details. I’ll quote from it below, but strongly recommend reading it in full.

Horgan mentions Compton’s 1959 interview with Pearl Buck, alongside an embarrassingly weird 1975 article in the Bulletin of the Atomic Scientists by J.C. Dudley, who speculates about a purported ether-like “neutrino sea” altering fusion rates, suggest a supercritical nuclear reactor might meltdown to the centre of the Earth and out the other side, and just about stops short of hypothesising mutant giant goldfish taking over the planet and eating everyone. He also puts on stilts Compton’s strange claim that calculations showed the risk of igniting the atmosphere from a nuclear bomb to be slightly less than three in a million, nonsensically taking this fictional risk figure to be independent for each bomb detonated, making the catastrophe inevitable given enough nuclear tests.

Bethe in 1975 The main point of Horgan’s post is to give what he says is a lightly edited transcript of a 1991 interview with Bethe, in which Bethe repeats his 1975 rebuttal of Dudley, where he reviews the arguments of Konopinski et al. and suggests that Buck must have completely misunderstood Compton. Bethe, in 1975, says “There was never any possibility of causing a thermodynamic chain reaction in the atmosphere […] ; it is simply impossible”.

Bethe in 1991 According to Horgan’s transcript, in 1991 he described the suggestion as “such absolute nonsense”. But, interestingly, he then took a very different line on Compton’s interview:

“Just to relieve the tension [at the Trinity test on July 16, 1945, Enrico] Fermi said, ‘Now, let’s make a bet whether the atmosphere will be set on fire by this test.’ [laughter] And I think maybe a few people took that bet. For instance, in Compton’s mind [the doomsday question] was not set to rest. He didn’t see my calculations [or] Konopinski’s much better calculations. So it was still spooking [Compton] when he gave [the interview to Pearl Buck in 1959]. “

So, no misunderstanding after all: Pearl Buck had accurately reported Compton, who was spooked in 1945, and still spooked even in 1959.

Bethe in 2000 By 2000, as I mentioned earlier, Bethe had a subtly different take again: “[T]he 1 in 300000 figure was made up by Compton ‘off the top of his head’, and is ‘far, far too large’.” So, now not “simply impossible”, merely “much, much more improbable than 1 in 300000” (whatever that was intended to mean).

Horgan’s blog has a very interesting commentary, which deserves quoting in full:

“Clarification from Alex Wellerstein, an Actual Expert: My friend and Stevens colleague Alex, an historian specializing in nuclear weapons, posted the following info-packed comment. Alex alludes to the possibility, also raised by Teller, that a nuclear explosion could ignite deuterium in the ocean. The term “Super” refers to a hydrogen fusion bomb.—John Horgan

John, I suspect that by the time you talked to him, especially after the (very silly) Dudley exchange, Bethe was pretty sick of this issue, and probably was downplaying the amount of effort and lingering concerns that existed in 1945. Dan Ellsberg, in The Doomsday Machine, makes an argument that they were more concerned than they let on later, and while I don’t totally think he is correct in his whole argument, I think he does a good job of showing that it wasn’t quite as dismissible as nonsense at the time, whatever Bethe thought of it.”

This sounds very plausible, with the downplay getting ever softer over the years. Bethe shifts from categorical denial in 1975 that anyone was worried, to accepting in 1991 that Compton was worried, not only at the time but even many years later, to a hint of a suggestion in 2000 that even he (Bethe) might have accepted that, even after the Konopinski-Teller calculations, the worries were not absolutely eliminated. What was Bethe’s true position? We may never know. If we can’t trust his 1975 account, and can’t quite trust his 1991 account either, then there’s certainly no reason to take his 2000 statement (by far the least considered of the three, a response to an email enquiry) as definitive. Perhaps – who knows? – the Bethe of 1945 wasn’t actually so far (in his subjective risk estimate) from the Compton of 1959?

But wait, back up! There’s something else extraordinary in Horgan’s interview. According to Bethe, Compton decided whether or not the 1945 Trinity test should go ahead without seeing any of the relevant calculations (Bethe’s or Konopinski’s). Of all the extraordinary claims around this issue, this is by far the strangest. The Konopinski-Marvin-Teller paper was written in 1946, but the substance of their arguments must have circulated before Trinity: that’s why everyone involved was aware both of the concern and that there were pretty strong reasons to think it wasn’t realistic. It’s not that hard to get at least the main ideas of the paper: Compton was unlikely to be deterred by the discussion of Compton scattering, for example. Compton was “spooked”, according to Bethe in that very interview. So what on Earth would have stopped him from looking at the calculations, or getting Bethe or Konopinski to explain them to him? And whose calculations, if not these, was Compton referring to when he said calculations showed the risk of igniting the atmosphere from a nuclear bomb to be slightly less than three in a million?

Bethe’s 1991 account makes absolutely no sense. Compton’s 1959 statement also makes absolutely no sense. Bethe pressed heavily and then more lightly on the downplay button over the years. Compton, so far as I’m aware, never elaborated further. Getting the history of catastrophic risk analyses straight seems peculiarly difficult.