Physicists in Italy claimed last week to have seen particles of dark matter. Their announcement has got their rivals riled and raises questions about what constitutes evidence of a new particle.

Rita Bernabei of the National Institute of Nuclear Physics in Rome presented her team's latest results on 16 April at an international meeting of particle physicists in Venice, Italy. Their detector, DAMA/LIBRA (Dark Matter Large Sodium Iodide Bulk for Rare Processes), located deep under the country's Gran Sasso mountain, seems to be observing dark matter, Bernabei says.

Most agree that the experiment is picking up something: “They're seeing a signal, there's no doubt about that,” says Tim Sumner of Imperial College London. But despite this, critics say that they don't believe the detector has found the elusive particles. “For me, it's not proof that they have seen dark matter,” says Gilles Gerbier, a physicist at the Centre for Atomic Energy in Saclay, France. He adds that he's stumped by what's causing the signal.

Dark matter is believed by most physicists and astronomers to make up some 85% of the matter in the Universe. Most theories predict that it is some form of massive particle that interacts very rarely — if at all — with regular matter such as atoms. To date, most believe that dark matter has been spotted only indirectly via its pull on rotating galaxies and its effect on the shape of the early cosmos.

It's not the first time that Bernabei's team has made this claim. In 2000, they also claimed to have directly observed dark matter. The team uses ultrapure sodium iodide crystals, which theory predicts will give off flashes of light when they are struck by dark-matter particles. After several years of collecting data in the late 1990s, the group saw an increase in the number of flashes every June and a decrease during December. The seasonal oscillation, they claimed, was due to Earth's annual motion with and against a stream of dark matter swirling around the Milky Way.

But rival groups failed to see the signal and heaped criticism on the group. They demanded proof that the Italians' detection was not caused by systematic errors such as natural radiation or glitches in their electronics.

This time around, the DAMA/LIBRA team seems to have addressed some of those concerns, according to Bernard Sadoulet, an astrophysicist at the University of California, Berkeley. For example, their new detector uses larger, less strongly radioactive crystals, and they have proven the stability of their detectors.

But once again, DAMA/LIBRA's rivals are coming up empty-handed. Sadoulet's own Cryogenic Dark Matter Search II detector announced a null result earlier this year (see Nature 452, 6–7; 2008). And other second- and third-generation experiments are similarly failing to see a signal. If the signal DARMA gets is dark matter, Sadoulet says, “it's certainly not the dark matter we were looking for”. He also continues to criticize the group for failing to share enough details of its data. “There has not been enough information provided to the community,” he says.

Bernabei defends her group's finding. Most other detectors try to measure direct collisions between a dark-matter particle and an atomic nucleus, she says. DAMA/LIBRA is the only experiment with the years of data needed to see the annual fluctuation in dark-matter particles. And the fluctuation matches “all the several requirements of dark matter”, she says, adding that two papers have just been posted on the popular preprint server ArXiv, which provide the details of her work.

But unless another team sees something, Gerbier says, it's unlikely that the wider physics community will accept the Italian claim. “There should be different ways of seeing the same thing,” he says. “A single experiment cannot discover dark matter.”

Bernabei agrees. Proving her team's findings will take “time and patience”, she says.