Zeroing in on the elusive neutrino's mass.

For an elementary particle that plays such crucial roles in processes ranging from radioactivity to supernova collapse, the neutrino has eluded characterization for a remarkably long time. As one step toward pinning it down, researchers have now obtained the best experimental estimate yet of a neutrino's mass.

But the relevance of this measurement, which sets an upper limit of 8 electronvolts on the mass of an electron antineutrino, remains clouded by concerns that an unknown physical effect may be interfering with the experiment. Researchers involved with this and other, related studies have detected puzzling anomalies in their data that are hard to explain using standard theory.

"Our result is correct only if there is no new physics involved," says Wolfgang Stoeff of the Lawrence Livermore (Calif.) National Laboratory, who heads the group that established the new limit. Stoeff reported the team's findings at last week's American Physical Society meeting in Washington, D.C.

To determine the neutrino mass, Stoeffl and his co-workers use a special apparatus to measure the energies of beta particles, or electrons, emitted by the decay of a radioactive form of hydrogen known as tritium. In such decay, one of the two neutrons in a tritium nucleus turns into a proton and sends off a particle-antiparticle pair -- an electron and an electron antineutrino.

By keeping track of the numbers of electrons detected at different energies, the researchers can plot an energy spectrum for tritium beta decay. They can deduce an upper limit on the neutrino's mass from the difference between the highest electron energy detected and the theoretical prediction of what that energy would be if the neutrino had no mass. That difference now stands at 8 electron-volts.

However, this tail end of the tritium beta-decay spectrum has a puzzling feature. When all the data in this region are taken into account, statistical measures suggest that the most likely value of the neutrino mass is a negative number -- something physically impossible.

"It's the opposite of what we were looking for," Stoeff says. "We have too many counts near the endpoint."

Other teams of researchers studying beta decay have consistently found similar small deviations from the expected number of electrons in this energy range. Moreover, the extraordinarily high precision of the Livermore experiment seems to preclude explanations that invoke energy losses or known atomic processes.

"It's had to explain," Stoeffl admits. "We have to find out exactly what it is."