Something is out there in the cosmos. We can't see it, we can't touch it and we know it's there only by the gravitational pull it exerts on cosmic objects. For decades the story of dark matter has been one revelation after another about what this mysterious material is not, a gradual winnowing of possibilities that has made physicists increasingly nervous. What happens when the last candidate gets crossed off the list? Will we be doomed never to glimpse the nature of the stuff that contributes about 25 percent of all mass in the universe?

This dreary narrative took a turn in a hopeful direction earlier this spring. Researchers uncovered one of the most intriguing clues in years: a hint of a new force that may allow dark matter to “talk” to itself. This insight would help explain what kind of particles dark matter might be made of.

he clue turned up in observations of a corner of the universe called the Abell 3827 cluster. Astronomers recently tracked dark matter's location within four colliding galaxies in this cluster by using a phenomenon known as gravitational lensing (the bending of light as it passes near massive objects). Observations made with the Hubble Space Telescope and the Very Large Telescope in Chile revealed that the dark matter surrounding at least one of the galaxies significantly lagged behind the ordinary matter there, suggesting dark matter particles were interacting with one another and slowing themselves down—a phenomenon never seen before.

Astronomers led by Richard Massey of Durham University in England surmise that because the interactions did not affect the normal matter, they must have occurred through some force other than gravity that influences only dark matter. An exchange of “dark photons” may create the force, for example. Such a situation potentially parallels the way regular protons interact with one another through the electromagnetic force: when two protons approach one another, each releases a photon—the force carrier of electromagnetism—and the other absorbs it. This exchange transfers momentum, causing both protons to separate.

The news has galvanized physicists in search of answers. “If this holds up, it is beyond a big deal,” says physicist Neal Weiner of New York University, who was not involved in the study. A scenario with dark photons is a change from the most basic and popular conception of dark matter as a single type of particle, commonly called a weakly interacting massive particle, or WIMP. But the idea that dark matter involves dark photons as well as exotic interactions might help explain some problems with the single-particle WIMP explanation for dark matter, such as why the centers of galaxies are less dense than expected.

This concept would also help physicists considerably narrow down the list of dark matter contenders. “Although we have evidence of dark matter from a huge variety of sources,” Weiner says, “we have so far no clear indication of anything other than its gravitational interaction. If it is shown to have self-interactions at this level, it will eliminate a huge number of models” for what dark matter could be. In particular, the finding, released online in April and published in June in Monthly Notices of the Royal Astronomical Society, may conflict with many popular versions of a hypothesis that dark matter is a particle predicted by supersymmetry theory. Supersymmetry—an appealing idea that attempts to explain many mysteries of physics, such as why the Higgs boson's mass is as low as it is—posits more particles in the universe than those that have so far been found. Yet if one of these particles (which could be a WIMP) were responsible for dark matter, most versions of the theory would not predict self-interactions.

The study's co-authors say it is too early to rule out a more mundane explanation for their observations. For instance, dark matter outside the colliding galaxies but along Earth's line of sight might be contributing to the gravitational lensing. “One caveat with this new study is that it's only one object,” says team member David Harvey of the Swiss Federal Institute of Technology in Lausanne. “There are unknown unknowns that may be changing the result.” And previous searches in other clusters have not seen signs of self-interacting dark matter, including a March Science study led by Harvey that analyzed 72 collisions of galaxy clusters rather than individual galaxies. Because clusters collide faster than galaxies, however, there is less time for dark matter to interact and drag behind, so the two findings are not contradictory.

If the recent observations turn out not to reflect new forces or dark matter interactions, Abell 3827 will become yet one more example of what dark matter isn't. Meanwhile searches for its particles in underground detectors continue to come up empty, and dark matter has so far failed to appear in CERN's Large Hadron Collider. Scientists hope these trends could soon change: the collider restarted in April at its highest energy levels yet, and the detectors are now extremely sensitive. “Dark matter has been so elusive, but we've never had the data we're going to have,” Harvey says. “I feel like it's now or never.”