Ever since astronomers realized that most of the matter in the universe is invisible, they have tried to sort out what that obscure stuff might be. But three decades of increasingly sophisticated searches have found no sign of dark matter, causing scientists to question some of their basic ideas about this elusive substance.

In October the most sensitive experiment looking for proof of the leading candidate for dark matter—theorized particles called WIMPs (weakly interacting massive particles)—reported null results, disappointing scientists once again. Now some researchers are reexamining dark matter candidates once written off as unlikely, and considering less satisfactory ideas such as the possibility that dark matter will turn out to be made of something more or less undetectable.

Physicists still have no proof that dark matter exists at all, but the evidence for it is substantial. The movements of stars and galaxies can apparently be explained only if there is much more gravitating matter in the universe than the visible stuff of atoms and molecules. Attempts to correct the discrepancy by rewriting the rules of gravity in Einstein's general theory of relativity have repeatedly failed.

WIMPs have long been the favored explanation for dark matter, in part because they could fit in with other popular ideas in physics such as supersymmetry—the suggestion that all known particles in the universe have as-yet-undiscovered partner particles. The lightest of these supersymmetric particles, it is thought, could be a WIMP, and could constitute the universe’s dark matter. Attempts to detect WIMPs during the rare occasions that they do bump into regular matter particles have been ongoing since the 1980s, but none have been successful. Most recently the Large Underground Xenon (LUX) project in South Dakota—by far the most sensitive search yet undertaken—reported that three months of data showed no signs whatsoever of dark matter (pdf).

If WIMPs do exist, they are running out of places to hide. "We are tightening the noose; we’re closing in," says LUX co-spokesman Richard Gaitskell of Brown University, who estimates that more than half of the possible WIMP models have already been disproved. The next five or 10 years should largely wrap up the WIMP search, either by discovering the particles at last or by essentially ruling out their existence. But if dark matter is not made of WIMPs, what is it?

Theoretical particles called axions are another oft-mentioned candidate. Much less massive than WIMPs and likely to interact even less frequently with regular matter, they are more difficult to search for—a fact that partly explains why only one major experiment is looking for axions today, whereas more than a dozen projects are hunting WIMPs. Yet axions, too, have a solid theoretical foundation and could easily explain an abundance of dark matter in the universe. "I don’t understand why axions tend to get ranked as number two," says Leslie Rosenberg of the University of Washington, who heads the Axion Dark Matter eXperiment (ADMX). "I would invert the order. But that’s my opinion." ADMX began in 1995, however, and has found no signs of axion dark matter so far. A more sensitive iteration of the experiment recently came online, and within three years the project will have either found axions or proved they do not exist, Rosenberg predicts.

With the searches for both WIMPs and axions nearing the finish line with no glimpses of success yet, more and more theorists are considering alternatives. "When people got nervous about WIMPs, other candidates came out of the woodwork," Rosenberg says. Some have suggested that a plenitude of small black holes throughout the universe could account for dark matter. Targeted astronomical searches have found no signs of such black holes, which should bend light from background objects in an effect called gravitational lensing. The idea could be resurrected with some theoretical finagling, however.

Another exotic possibility attracting increased interest is quark matter—an extremely dense phase of matter made of strange quarks (exotic cousins of the up and down quarks that form protons and neutrons). Quark matter could be created inside very massive neutron stars, and in sufficient quantities it could make up a population of quark stars that would emit no light but could exert a gravitational pull on normal matter.

Those are just a few of the ideas populating the wide-open landscape of potential explanations for dark matter. "I doubt we’ve thought through all the interesting possibilities," says theorist Matt Strassler, a visiting physicist at Harvard University. "We may get lucky" and find the answer soon, he says, "or this may drag on for 100 years or more."

The scariest possibility may be that dark matter is made of something impossible to find—some particle that interacts with regular matter only via gravity and no other force. In such a case researchers would have no hope of catching it in a detector. "If we move into a mode where our most favored particles are simply not detectable, we have the classic scientific challenge, which is how do you verify such a theory?" Gaitskell asks. "At that point you’re almost a failure—you have a theory that’s almost impossible to test."

Perhaps thankfully, such a particle seems somewhat unlikely from a theoretical standpoint. "With a particle that really only interacts gravitationally, you have to ask how it ever got produced in the universe in the first place," says Stanford University theoretical physicist Peter Graham. "Axions and WIMPs both have very nice natural production mechanisms and reasons why they might be here in such abundance. That's what disfavors these other models." Still, they remain on the table.

Even if physicists cannot detect dark particles directly, they hold out the hope that they might find indirect evidence of dark matter with particle accelerators such as the Large Hadron Collider in Europe. The LHC could produce WIMPs or other dark matter candidates when it smashes protons together in powerful collisions. Another hope is that astrophysical signals, such as gamma-ray light from the center of the galaxy, might reveal the presence of dark matter particles if dark matter annihilates itself when two particles make contact. Some hints of such signals have been claimed, but they are far from definitive.

Ultimately, most physicists in the hunt say they do not care whether their particular conception of dark matter turns out to be right, as long as they eventually get an answer. "As in all research, there is never a guarantee of success. All we can do is to continue to attempt to answer the most important science questions," says Blas Cabrera of Stanford University, who leads the Cryogenic Dark Matter Search (CDMS), another WIMP-detection experiment. Lately, more physicists are facing the specter of possible failure. "I would be horrifically disappointed if we didn't discover dark matter," Rosenberg says. "It is within our grasp, and I really want to know what it is."