Dark influence: most of the universe's matter is out of sight, but not out of mind.

Imagine trying to figure out what's happening in a film from just a few scattered frames near the end. Now, let's make it even more challenging: The star of the movie is invisible as you watch the rest of the characters going about their business. That's what astronomers are up against in their quest to understand the cosmos.

The structures of the universe evolve so slowly from the human point of view that they appear as still images. To make matters worse, astronomers in the 1970s began to realize that most of the material in the universe is unseen. They called it dark matter. Just as blowing leaves suggest the presence of wind, the motions of visible matter such as stars and galaxies betray the gravitational pull of dark matter.

Without dark matter, the universe would look totally different. Vast clouds of dark matter draw in gas from their surroundings, concentrating it into galaxies containing billions of stars. Without galaxies, there would be none of the cosmos' visible structures--no stars, no planets, and no people wondering what it's all about.

By determining the distribution of dark matter in the modern universe--where dark matter outweighs visible matter by factor of six or so--astronomers can work backward to puzzle out how it has behaved since the Big Bang. They have already gleaned some precious pieces of information about dark matter this way and are now looking for more details.

On cosmic size scales ranging from individual galaxies to the whole shebang, the distribution of dark matter provides clues about this substance's nature and past. So far, astronomers have uncovered more about the large-scale organization of dark matter (SN: 1/5/02, p. 5) than about its distribution on the scales of galaxy clusters and individual galaxies.

Now, the most precise observations ever made of individual galaxies are revealing vexing details about the location of dark matter in individual galaxies. Some unknown effect appears to be clearing dark matter out of the centers of galaxies. And new evidence indicates that when speeding fragments of dark matter meet, they don't collide as other matter does but pass right through each other, ghostlike.

WHAT WAGS THE DOG? Like all substances that have mass, dark matter is gregarious. Gravitational attraction gathers it into clumps called dark matter halos. Because dark matter doesn't appear to be affected by any three other than gravity, computer simulations indicate that all the halos should be similar--roughly spherical and much denser at their cores than at their edges.

Astronomers can probe the dark matter halos around individual galaxies by charting the motion of different, onionlike layers of stars and gas that make up the galaxies. The faster a particular layer of stars and gas rotates, the more matter there must be between it and the galaxy's core. Otherwise, the matter would fly off into deep space. However, if only stars, gas clouds, and other visible matter were supplying the gravity, there wouldn't be enough to hold together galaxies. So, astronomers infer how a galaxy's dark matter is distributed by looking at the apparent gravitational short fall in each onion layer.

Previous telescope studies suggested that halos have a variety of structures, but astronomers have debated whether these observations have enough resolution to rule out the expected symmetry.

Now, having obtained the highest-resolution observations yet of motions within galaxies, Joshua D. Simon of the University of California, Berkeley and his colleagues seem to have settled the question.

The researchers used a radio telescope array in Hat Creek, Calif., and a visible-light telescope at Kitt Peak, Ariz., to track gas clouds as they orbited the centers of five galaxies. By measuring the clouds' speeds and factoring out the influence of visible matter, the scientists measured the density of the dark matter halos at various distances from their centers. One halo was densely packed toward the galaxy's center, as simulations predicted, another had a density that was the same everywhere, and the density profiles of the other three were somewhere in between these extremes.

Astronomers aren't sure how to explain this variety. One possibility is that dark matter responds to forces other than gravity. If dark matter particles can collide with each other, for example, then they might avoid getting crowded together near the centers of galaxies, despite gravitational attraction. Another possibility is that ordinary matter, the kind that emits or absorbs light, can somehow alter the distribution of the dark matter.

"Right now, the simulations that people are running only include dark matter," says Simon. "They don't include stars and gas, which are clearly a major component of galaxies."

Julio F. Navarro of the University of Victoria in British Columbia suspects that interactions between ordinary matter and dark matter underlie the variety of halo structures. "There may be complex interactions ... that may lead to significant changes in the properties of the dark halo compared to the ones we would get if there was no galaxy there," he says.

Many galaxies, perhaps even our own, have a rotating bar of stars and gas at the center. Some theorists have suggested that as such bars sweep away dark matter, they could thin it to different extents. That might account for the range of dark matter distributions seen by Simon's group and others.

Dark matter was dreamed up as something that would push ordinary matter into large-scale structures in the universe. If the sweeping-bar scenario is right, then the tables are turned, at least near the centers of galaxies, where ordinary matter is calling the shots and pushing around dark matter.

MISSING HALOS The variety of halo structures isn't the only problem that makes some astronomers wonder whether there's real substance to the notion of dark matter. For years, astronomers have scratched their heads over a second problem: a shortage of the smallest dark matter halos.

"Despite its spectacular successes, [the standard theory of] dark matter has had these two big problems," says Priyamvada Natarajan of Yale University.

Astronomers have detected halos of different sizes, and they suspect that the larger halos form through the mergers of smaller ones. Computer simulations of this process match the observed range of sizes, with the exception of the smallest halos within which tiny galaxies, called dwarf galaxies, form. In vast orbits around our own medium-size galaxy, astronomers have found about a dozen of these dwarfs, but simulations of dark matter predict there should be 50 or so dwarfs around the Milky Way.

To address this problem, David Spergel of Princeton University has proposed that pieces of dark matter ricochet off one another as billiard balls do. Previously, astronomers assumed that dark matter particles interacted only via gravity and that they passed through one another with little consequence.

If dark matter particles can collide, then dark matter halos would tend to merge more frequently than they would otherwise. So, by the modern period of the universe, most of the smaller halos would already have merged into larger ones, and that would explain the apparent dearth of small halos.

If dark matter particles can collide and interact, these processes should be reflected in the distribution of dark matter on larger scales. Dense clusters containing hundreds or thousands of galaxies would be surrounded by a very large dark matter halo, in which smaller halos associated with each galaxy would be embedded. In such a cluster, collisions between dark matter particles should eventually erase boundaries demarcating the smaller halos.

To investigate this scenario, Natarajan's team mapped dark matter in several massive galaxy clusters. The team exploited a cosmic optical effect known as gravitational lensing.

As a consequence of relativity, the gravity of massive objects distorts the fabric of space-time and thereby the pathways of light rays passing the objects. The amount of this bending depends on the mass of the object. By measuring the bending and having a measure of how much visible matter is in the object, scientists can infer how much dark matter must also be present in the object.

Using Hubble Space Telescope images taken for earlier studies, Nataraj an looked at how relatively nearby clusters of galaxies bend the light of distant galaxies. By subtracting the lensing effect of the ordinary matter, Natarajan's team zeroed in on the portion of lensing due to dark matter. In this way, the team mapped out, in unprecedented detail, where the dark matter lies in these clusters.

"No one's ever been able to do this kind of detailed, high-resolution study" of dark matter distribution in clusters, she says. The analysis revealed lots of galaxy-size clumps of dark matter within the overall cluster halo, so the boundaries haven't vanished.

This finding rules out the idea that dark matter particles can collide and interact with one another, Natarajan contends.

Spergel counters that the observations don't preclude interaction among dark matter particles but do put limits on it.

"[Natarajan's] work shows that there is no evidence for interactions strong enough to affect dark matter on the cluster scale," says Spergel. But it doesn't rule out interactions--other than gravitational effects--among dark matter particles colliding at low speeds. After all, he notes, some of the more familiar particles in nature, such as neutrons, collide more easily at low speeds.

Natarajan raises another possibility. Perhaps the dwarf halos are out there, she says, but have escaped detection because, for some reason, they aren't associated with detectable ordinary matter such as galaxies (SN: 2/26/05, p. 131).

What dark matter is actually made of is still anyone's guess. "Fundamentally, we know very little about the nature of the dark matter," Spergel admits.

The most popular idea among astronomers is that it is composed of subatomic particles with exotic properties. One property that dark matter shares with ordinary matter--mass--provides at least a toehold for experimentalists. A particle accelerator might someday produce dark matter particles and finally settle the mystery of their identity.

Until a positive identification is made, Spergel says, increasingly detailed observations of galaxies and galaxy clusters offer the best approach for unveiling the invisible stuff that is everywhere.