I have erased several comments that only increased the amount of confusion and changed the filename to temporarily break links from crackpots' blogs. I apologize to everyone who was affected by any confusion in previous versions of this article, and I hope that the text below is now more or less OK and perhaps the clearest report on the Internet:
See my 1998 notes that contain almost everything you need to know about neutrinos: not much has changed for 9 years although people different from your humble correspondent were less certain about most of these things than I wasEvidence for several types of neutrino oscillations has been known for a decade or more. That includes atmospheric neutrino oscillations (in which muon neutrinos get "lost"), solar neutrino oscillations (in which electron neutrinos get "lost"), and a lab experiment called LSND in Los Alamos.
A simple oscillation in between two neutrino flavors - electron neutrino and muon neutrino - was a natural candidate theory to explain the observations but it couldn't agree with details of the LSND data which is why the LSND results were questioned. The next-to-simplest seemingly natural candidate model was a two-flavor oscillation that would include a sterile neutrino, a new kind of "fourth" neutrino without a charged partner.
Mainstream physicists like myself have always considered sterile neutrinos to be an unjustified construct. Such a particle never emerges from string theory - it seems fair to say that string theory predicts that there can't be any sterile neutrinos - but even outside string theory, models with such a particle feel unnatural. If you add such a new spinor field - which is ugly - it is expected to get a huge (GUT) mass anyway. There's no reason to expect a light sterile neutrino.
Today, Fermilab's MiniBooNE experiment has confirmed that the simplest model suggested by the "anomalous" LSND signal was incorrect and a more subtle explanation than the simple two-flavor oscillation is necessary: the reason is, if I simplify a bit, that they didn't observe any electron neutrinos. You can now view the weak LSND signal as a fluke if not an error. The message is: forget about the the LSND signal.
The result therefore rules out the possibility that the observed oscillation from LSND is a two-flavor oscillation involving a new sterile neutrino. See
The only thing that seems clear to me is that the phenomena studied by LSND and MiniBooNE can't be explained by a simple oscillation between two eigenstates if you measure things properly. They have shown that the most drastic conclusions from the LSND results are wrong but they don't seem to offer any real conclusions except for refuting the significance of the previous LSND signal: nothing has changed about this observation for several hours.
There are too many uncertainties here which is why titles like "Fermilab experiment resolves long-standing neutrino question" seem inappropriate to me.
There will be many talks about their results today and tomorrow - including a talk at Harvard - and the situation will surely clear up soon. My guess is that the three-flavor oscillation as described by the Standard Model simply works fine - they are just unable to say it clearly - and any evidence of new physics suggested by LSND was flawed. There is no good reason whatsoever to expect any new physics in the neutrino oscillations except for the trivial fact of the existence of these oscillations which sane physicists have known to be a fact for a decade or more.
While I think it's obvious that their results are a big setback for all proponents of sterile neutrinos, they don't hesitate to propose a new experiment - LENS - with a special focus to study possible sterile neutrinos. See PhysORG.com. Sorry to say but it reminds me of the Woit-Smolin type of "falsifiable" pseudoscience whose goal is to generate stupid predictions and test them - while paying tens of millions of dollars - and never care about the results of the experiments.
Science, as I understand it, is about trying to find out how the world actually works, instead of creating unlikely fairy-tales how it could work with the goal of getting predictions out of these fairy-tales that are as spectacular and easily visible as you can get. ;-)
When you generate silly theories and make and test silly predictions, critics of science like Woit and Smolin will confirm that you're doing good science. Sorry, I won't. Scientists must actually care what has already been measured if it has already been settled. Their opinions must be influenced by results of careful experiments, calculations, and arguments.
In this sense, I think that designing expensive experiments to test sterile neutrino models is a bad experimental science simply because sterile neutrino theories are bad theoretical physics and it is very unlikely that such an expensive experiment will bring at least something interesting. Today, LSND seems as a controversial investment that only brought a fake signal while MiniBooNE could only show that the LSND signal wasn't real but it couldn't say anything else. I hope that LENS will say at least something else even in the likely case that sterile neutrinos don't exist. Otherwise I think it would be even worse an investment than LSND and MiniBooNE because it would be a repeated bad investment.
Moreover, I really dislike their comments that is should now become really exciting to test the sterile neutrinos even more accurately. Sterile neutrinos are silly and the news today have strengthened this statement. An honest scientist will take these observations into account instead of designing ever more unnatural theories with sterile neutrinos and ever more expensive experiments to test these bad theories. If funding concerns drive them into twisting reports about their results, it's bad and it remotely resembles climate science.
And that's the memo.