According to the prevailing "cold dark matter" theory of the evolution of the universe, every galaxy is surrounded by a halo of dark matter that can only be detected indirectly by observing its gravitational effects. This theory faced a challenge in 2003, when a team of astronomers reported a surprising absence of dark matter in elliptical galaxies. But a new analysis published in the September 29 issue of the journal Nature provides an explanation for the earlier observations that fits comfortably with the standard theory and puts the dark matter back into elliptical galaxies.

"These are very normal, nearby elliptical galaxies that they studied, and if those galaxies don't have dark matter it calls into question the whole theory of cold dark matter," said Joel Primack, professor of physics at the University of California, Santa Cruz, and a coauthor of the Nature paper.

"A dearth of dark matter in elliptical galaxies is especially puzzling in the context of the standard theory of galaxy formation, which assumes that ellipticals originate from mergers of disk galaxies," added Avishai Dekel, professor of physics at the Hebrew University of Jerusalem and first author of the Nature paper.

"Massive dark matter halos are clearly detected in disk galaxies, so where did they disappear to during the mergers?" said Dekel, currently a visiting researcher at UCSC.

Primack, one of the originators and developers of the cold dark matter theory, uses supercomputers to run simulations of galaxy formation and the evolution of structure in the universe. The new paper used simulations of galaxy mergers run last year by Thomas J. Cox, then a graduate student working with Primack at UCSC and now a postdoctoral researcher at Harvard University.

The simulations show that the observations reported in 2003 are a predictable consequence of the violent galactic mergers that give rise to elliptical galaxies, Primack said. The simulations were analyzed by Dekel, Felix Stoehr, and Gary Mamon at the Institute of Astrophysics in Paris, where Dekel holds a Blaise Pascal International Chair. UCSC graduate student Greg Novak also contributed to the analysis.

Elliptical galaxies are thought to form when two spiral galaxies collide and merge. Whereas spiral galaxies are dominated by flattened, rotating disks of stars and gas, elliptical galaxies are round, smooth collections of stars.

Evidence for dark matter halos around spiral galaxies comes from studying the circular motions of stars in these galaxies. Because most of the visible mass in a galaxy is concentrated in the central region, stars at great distances from the center would be expected to move more slowly than stars closer in. Instead, careful observations of spiral galaxies show that the rotational speed of stars in the outskirts of the disk remains constant as far out as astronomers can measure it.

The reason for this, according to cold dark matter theory, is the presence of an enormous halo of unseen dark matter surrounding the galaxy and exerting its gravitational influence on the stars. Additional support for dark matter halos has come from a variety of other observations.

In elliptical galaxies, however, it has been difficult to study the motions of stars at great distances from the center. The 2003 study (A. J. Romanowsky et al., Science 301:1696-1698) focused on bright planetary nebulas in the outer parts of four nearby elliptical galaxies. Planetary nebulas are old stars that have blown off their outer layers and glow brightly in characteristic wavelengths of light. The researchers were able to determine the line-of-sight velocities of large numbers of planetary nebulas in these elliptical galaxies. They found a decrease in the velocities with increasing distance from the center of the galaxy, which is inconsistent with simple models of the gravitational effects of dark matter halos.

Part of the explanation put forth in the new Nature paper lies in the fact that the velocities were measured along the line of sight. "You cannot measure the absolute speeds of the stars, but you can measure their relative speeds along the line of sight, because if a star is moving toward us its light is shifted to shorter wavelengths, and if it is moving away from us its light is shifted to longer wavelengths," Primack explained.

This limitation would not be a problem if the orbits of the observed stars were randomly oriented with respect to the line of sight, because any differences resulting from the orientations of the orbits would average out over a large number of observations. According to Cox's simulations, however, the stars farthest from the center of the galaxy at any given time are likely to be moving in elongated, eccentric orbits such that most of their motion is perpendicular to the line of sight. Therefore, they could be moving at high velocities without exhibiting much motion toward or away from the observers.

To understand why, it is necessary to look at what happens to the stars during galaxy mergers. As the merging galaxies interact, the stars themselves do not collide because they are separated by great distances, so the two galaxies essentially pass through one another. But the huge gravitational fields of the galaxies cause powerful tidal disturbances. Some of the stars are flung outward in extended tidal tails as the cores of the galaxies pass close by one another and spin apart. Sometimes the cores remain connected by a tidal bridge of stars and gas. Eventually, gravity pulls the cores back together, and the stars that were flung outward fall back in toward the center.

"In the merger process that produces these galaxies, a lot of the stars get flung out to fairly large distances, and they end up in highly elongated orbits that take them far away and then back in close to the center," Primack said.

To an observer outside the galaxy, a star on such an elongated orbit would only appear to be far from the galactic center if the long axis of its orbit is more or less perpendicular to the observer's line of sight. If the long axis of the orbit is aligned with the line of sight, the star would always appear to be in the crowded center of the galaxy from the perspective of the observer.

"If we see a star at a large distance from the center of the galaxy, that star is going to be mostly moving either away from the center or back toward the center. Almost certainly, most of its motion is perpendicular to our line of sight," Primack said.

The simulated mergers involved typical spiral galaxies, each embedded in a halo of cold dark matter. The simulations followed the gravitational and hydrodynamic evolution of the merger systems, taking into account the complicated feedbacks from star formation, supernovae, and the heating and cooling of gases in the galaxies. Each simulation was then "observed" from three different directions and at two slightly different times after the merger.

From more than 200 merger simulations run by Cox on a supercomputer at UCSC, the researchers analyzed 10 mergers that yielded elliptical galaxies with masses similar to those of the galaxies observed in 2003. The results were completely consistent with the reported observations, Primack said.

"Our conclusion is that what they saw is exactly what the cold dark matter model would predict," he said. "Their data are great, and this actually gives us more insight into how elliptical galaxies form."

"We predict that other velocity tracers in the same elliptical galaxies will show higher velocities if they are less concentrated toward the galaxy center or if they move on more circular orbits," Dekel said. "This is likely to be the case for compact star clusters, which are also observable in the outskirts of elliptical galaxies."