Is finding a missing flavor the same as finding a new one?

Disappearance means fewer events than expected

Suppose a source makes mostly muon neutrinos, and a distant detector sees fewer muon-flavor interactions than expected. That is a disappearance measurement. It tells you how likely the original flavor is to be detected after the trip.

The word means a shortfall against a prediction. That prediction has to include the source’s intensity and spectrum, the geometry, the interaction probabilities and the detector efficiency. A wrong flux estimate or an unmodeled loss of efficiency can also lower the count.

Near detectors, control samples and the way the shortfall changes with energy or distance help tell these causes apart. Super-K compared atmospheric neutrinos over different path lengths; reactor experiments compare near and far detectors. These are different kinds of control. A shortfall that follows a pattern says more than a single missing amount.

Sources: [1] · [2] · [3] · [4]

Appearance means more events than other processes can make

Now look in the far sample for electron-flavor interactions. More of them than predicted without oscillation can be evidence of appearance. But a beam of mostly muon neutrinos may already contain some electron neutrinos when it leaves the source.

That contamination has to be estimated, and neutral-current events or other particles can mimic an electron-like signal. Near-detector measurements and calibration pin down those contributions. The analysis also needs to know what fraction of real electron interactions make it into the selected sample.

In the standard picture of three active flavors with unitary mixing, the probabilities of ending in each flavor add up to one. Detected counts do not add up that simply, because each flavor has its own interaction threshold, cross section and selection efficiency.

Sources: [4] · [3] · [5]

T2K looked for electrons in a muon-neutrino beam

T2K sent a beam of mostly muon neutrinos from J-PARC to Super-Kamiokande, about 295 km away. Electron-like Cherenkov events in the far detector were the appearance signal; near-detector measurements pinned down the beam and the interaction model.

The analysis, submitted in 2013 and published in 2014, selected 28 electron-neutrino candidates. With its fitted background and oscillation assumptions, it excluded zero electron-neutrino appearance at 7.3 standard deviations. These numbers belong to that analysis; they do not summarize later T2K results.

The result rested on energy information, event selection and systematic uncertainties, not just on seeing some electrons. Electron contamination and misidentified events were already part of the no-appearance expectation; only the excess needed explaining.

Sources: [3] · [5] · [6]

OPERA asked whether tau flavor had arrived

OPERA studied the muon-neutrino beam sent from CERN to Gran Sasso, roughly 730 km away. Making a tau requires an energetic enough interaction, and the short-lived tau had to be recognized from its decay products in fine-grained emulsion.

Its 2015 appearance paper reported a fifth tau-neutrino candidate and a combined rejection of the background-only hypothesis at 5.1 standard deviations. Charm decays and secondary hadron interactions can mimic parts of the event shape, so the analysis used the decay geometry, the kinematics and backgrounds estimated channel by channel.

DONUT had earlier detected tau-neutrino interactions directly, showing that tau flavor could be seen in a suitable source sample. OPERA asked a different question: does tau flavor appear after a beam of mostly muon flavor has traveled? Similar particle identification answered different questions.

Sources: [7] · [8] · [9]

Together, the two measurements are harder to fool

A shortfall in the original flavor and an excess in another are vulnerable to different experimental mistakes. When both agree with one propagation model across energies and distances, the case is stronger than repeating the same kind of count in one instrument.

The source and detector data are statistical samples, and a physical model of flavor probabilities connects them. No single neutrino is identified at the source, followed all the way and recognized again at the far detector.

A disappearance does not guarantee that the new flavor will be seen. If the arriving flavor is below its production threshold or hard to identify, the detector may miss it. For any result, ask the same two questions: which flavor-sensitive quantity was measured, and which alternatives went into the expectation?

Sources: [4] · [10] · [3] · [7]

Try it in the Lab

Primary sources & revision

1. Super-Kamiokande Collaboration · Evidence for Oscillation of Atmospheric Neutrinos (1998)
2. Daya Bay Collaboration · Observation of electron-antineutrino disappearance (2012)
3. T2K Collaboration · Electron-neutrino appearance in a muon-neutrino beam (2014)
4. Particle Data Group · Neutrino Masses, Mixing, and Oscillations (2024)
5. T2K · Original electron-neutrino appearance publication
6. T2K Collaboration · The T2K Experiment (2011)
7. OPERA Collaboration · Tau-neutrino appearance in the CNGS beam (2015)
8. INFN · Fifth tau neutrino detected by OPERA (2015)
9. DONUT Collaboration · Observation of Tau Neutrino Interactions (2000/2001)
10. Nobel Committee · Neutrino oscillations, scientific background (2015)

First published 2026-10-09; last revised 2026-10-09. Original explanatory text and diagrams by Sandbox Physics. Illustrations are schematic; no experimental event records are reproduced here. This is an independent educational publication, not an official Nobel or experiment collaboration publication.