Bernd T. Matthias and the green flash

While searching for Web pages that contain the phrase “green flash”, I came across a page posted by the UCSD Library's Mandeville Department of Special Collections, referring to the manuscripts of Bernd T. Matthias. Matthias was well known for his work on superconductivity, not optics. I knew that he and John H. Taylor, a psychologist specializing in vision and the psychophysics of color, had published a paper on a green flash they had observed from the NASA Convair 990 flying observatory, in 1969.

The library's Web page mentioned a green-flash proposal, as well as the Nature  paper. As UCSD is less than 30 km from here, I thought I'd go check this out. Currently, parking is free on weekends; and the Special Collections department is open Saturday from 10 to 1. After an early lunch, I was there by noon.

I found the place very elegantly laid out; they even supply paper and pencils to take notes with. A page brought out the two boxes containing Matthias's green-flash papers, and I read through them.

The first box, dealing with the Nature  paper, documents what I had learned from Matthias's former secretary: that Taylor and Matthias had sent the original negatives of their green-flash photos to Nature  along with their manuscript, and that Nature  had subsequently managed to lose the negatives. On the one hand, that's another reason not to publish in Nature ; on the other, I think Taylor and Matthias were foolish to send off such important originals with their manuscript. No professional photographer would make this mistake.

[NOTE: a couple of people have pointed out that journals often want the original negatives, as later-generation copies usually have lower quality. However, Taylor and Matthias could at least have made backup copies to keep. I continue to think they should have been more careful with their unique material to protect it against loss.]

The second box is more interesting. It explains why these people were working so far from their areas of expertise. One of them (my guess is, Matthias) had somehow got the notion that the green flash was connected with stimulated-emission phenomena in the upper atmosphere; the folder contained some copies of papers on aurorae and theoretical work on stimulated Rayleigh scattering. The other author (probably Taylor, who was a member of the Optical Society) adhered to the conventional view that peculiarities of atmospheric refraction would suffice.

Evidently they had argued about this matter, and had decided they could settle it by taking spectra of green flashes from the NASA flying observatory. So they threw together a proposal, asking for thousands of dollars in both salary support and special equipment, to try it. The proposal contained no quantitative calculations to support the idea; it was just arm-waving. Although they claimed that the observations would allow them to decide which mechanism was the right one, they never explained just how this would be done. The only references (besides their Nature  paper) in the proposal that had anything to do with green flashes were Minnaert's book, and O'Connell's — and Jacobsen's awful paper in JRASC, which Minnaert had directed them to in a letter (included in the first box). [Minnaert had said in his letter that this observation should be repeated; that was basically the point of their proposal.]

Based merely on Minnaert's second-hand account of Dijkwel's spectral observations of green flashes, and (probably) on O'Connell's mistaken assertion that green flashes were caused by the upper atmosphere, these guys went off on a tangent into far left field, without bothering to look up the original of Dijkwel's paper, or even the green-rim spectra of Danjon and Rougier cited by O'Connell. They made noises about nitrogen emissions, and claimed that color photographs of green flashes showed only yellow and green and therefore (!) did not support the idea of a continuous spectrum. (A common misunderstanding about color film is involved here.)

Well, this is sloppy science of the worst sort. As I point out in the reading page, Minnaert's account is almost entirely second-hand; but one of the few references he actually gives is to Dijkwel's paper. And as long as O'Connell gives over 300 references, including the works of Danjon and Rougier, you'd think any responsible person who was serious about the subject would at least try to look some of them up. I suspect that, like many other people, these folks were just not taking green flashes seriously.

Not surprisingly, both NASA Ames and NASA Headquarters rejected the proposal; it restores one's faith in the refereeing system to read the terse letters of rejection. Astonishingly, Matthias and Taylor then tried first ONR and then the Research Corporation, with equally negative results. The subsequent versions of the proposal show no learning curve whatsoever: there is no indication that they ever bothered to look at the primary GF literature. To their credit, Taylor and Matthias at least did not try to publish the “stimulated emission” idea, but confined it to their proposal.

Obviously, there's a moral here. No matter how good you are (and Matthias was good enough to have been nominated for a Nobel prize), you can't get by on just your reputation alone. Some basic scholarship, such as the dog-work of digging through references in the library, is necessary to bring any project to fruition.

There's another sub-text to be pointed out. Everybody makes mistakes. Galileo had a crank theory about the motion of the Earth and the cause of the tides (he thought the Moon had nothing to do with them). Newton had a crank theory about the relative distances of stars of various magnitudes; it involved denying a theorem, proved in Classical antiquity, about the maximum number (12) of identical spheres that can touch a central sphere of the same size. (Newton thought it might be 13, or even 14.) Name any great scientist, and someone can point out a serious error that person has made.

What makes science work, as a joint effort, is that one person's error can be discovered and corrected by someone else. Eventually, the truth wins out (though sometimes it takes a long time). Eventually, the rest of us are left scratching our heads, and wondering how anyone so smart could have done anything so stupid. Errare humanum est. We're all human.

Sometimes it's good to remind yourself that you, too, can make mistakes. Worse, you can make mistakes and not recognize them, even when they're pointed out.

If you find I've made some mistakes on these pages, please let me know. I hope I'll get them corrected quickly.

Taylor and Matthias: Chapter Two

Amazingly enough, that isn't the end of this story. Early in 1999, I was asked to talk on green flashes to the local Optical Society section. One of the people in the audience was Roswell Austin, one of the Vis Lab's old-timers. He told me he had actually been on the CV-990 flight when Taylor and Matthias took their green-flash pictures.

What's more, he had actually saved some reprints of their paper when the Vis Lab closed down. So, some time later, I found myself the proud owner of an actual copy of the reprint. The color plate of their green flash is unusual for Nature at the time, when color work was rare and even more expensive than it is today.

Thanks, Ros!

© 1999, 2001, 2003, 2005 – 2009 Andrew T. Young

Back to the GF home page
Back to the library adventures page