Gamma-ray bursts, with their ability to pierce through gas and dust to shine brightly across the universe, are revealing areas of intense star formation and stellar death where astronomers have been unable to look - the dusty corners of otherwise dust-free galaxies.

The conclusion comes from a survey of "dark" gamma-ray bursts - bright in gamma- and X-ray emissions, but with little or no visible light - reported at a meeting of the American Astronomical Society in Pasadena, Calif., by astronomers from the University of California, Berkeley, and institutions around the world.

"Our study provides compelling evidence that a large fraction of star formation in the universe is hidden by dust in galaxies that do not appear otherwise dusty," said Joshua Bloom, associate professor of astronomy at UC Berkeley and senior author of the study.

Star formation occurs in dense clouds that quickly fill with dust as the most massive stars rapidly age and explode, spewing newly created elements into the interstellar medium to seed new star formation. Hence, astronomers presume that a large amount of star formation is occurring in dust-filled galaxies, although actually measuring how much dust this process has built up in the most distant galaxies has proved extremely challenging.

Long-duration gamma-ray bursts, the most brilliant flashes of light in the universe, are thought to originate from the explosion of massive stars. These events create two pencil-like beams of light, akin to lighthouse beacons, bright enough to be seen from as far away as 13 billion light years, near the limits of the observable universe.

While most gamma-ray bursts continue to shine brightly in optical light for many hours after the gamma-ray emission subsides - a phenomenon known as an 'afterglow' - those with little or no detectable afterglow, dubbed "dark GRBs," have puzzled astronomers. Some have speculated that most were so far away, and thus at such high redshift, that their optical afterglow shifted out of the wavelength region that optical telescopes can detect.

Redshift refers to the Doppler-shifted reddening of light from distant stars because they are speeding away from us, a consequence of the expansion of the universe after the Big Bang.

"Whatever the cause, it was like hearing the foghorn without seeing the lighthouse," explained Bloom. "Something interesting was happening towards those shores."

The new study, which focused on 14 bursts whose optical light was either much fainter than expected or completely absent, shows that almost every "dark" gamma-ray burst has a host galaxy detectable with Earth's largest optical telescopes - in this case, the Keck 10-meter telescopes in Hawaii. Because these galaxies would not be detectable if they were at high redshifts, this indicates that most "dark" bursts are similar to normal bursts with an afterglow, except that nearly all of the visible light is obscured by patchy dust within these host galaxies.

The findings suggest that gamma-ray bursts may be able to help track the rate at which stars form and die in distant galaxies, and confirm previous estimates that "25 percent of the time, when massive stars form, they form in a dusty place," said UC Berkeley graduate student Daniel Perley and lead author of the study.

"However, based on our survey of these dark gamma-ray bursts, the galaxies look normal and not dust filled," he said. "The dust is probably in clouds and knots around the forming stars."

Bloom and Perley were using some of the world's largest telescopes, the twin 10-meter telescopes of the W. M. Keck Observatory, to look for the host galaxies of "dark" gamma-ray bursts when Cenko, recently arrived from Palomar Observatory, suggested focusing on a specific sample of bursts observed by Palomar's 60-inch telescope.

Through March 2008, Palomar conducted follow-up observations of 29 bursts discovered by NASA's Swift gamma-ray satellite, 14 of which were classified as dark. The Swift mission, equipped with a gamma-ray detector and X-ray, ultraviolet, and optical telescopes, is operated by NASA's Goddard Space Flight Center.

For 11 of these 14 dark bursts, the team successfully detected a distant galaxy hosting the explosion, while the remaining three bursts without detectable hosts had faint optical counterparts.

This indicates that none of these bursts had come from the most distant regions of the universe, since at distances greater than about 12.9 billion light years all the detectable light from both the afterglow and the host galaxy would be shifted into the infrared due to the expansion of the universe.

"And while 12.9 billion light years is a large distance even by most astronomers' standards, gamma-ray bursts are so powerful that if these were frequent occurrences 13 billion years ago, we ought to be detecting large numbers of those same explosions today as high redshift events," Cenko said.

"We don't, which indicates that the first stars formed at a less frenzied pace than some models suggested."

The lack of any very high redshift events in the sample indicates that these distant explosions cannot comprise more than a few percent of all gamma-ray bursts, Cenko said. However, such distant bursts are known to exist. Just two months ago, a gamma-ray burst at a distance of 13.1 billion years was discovered.

"Putting this recent event together with the others in our study, for the first time we can provide both an upper and lower limit to the fraction of gamma-ray bursts at very high redshift," Perley said. Specifically, the authors conclude that the high redshift fraction is between 0.2 and 7 percent.

Because none of the 14 bursts studied in the survey is at this distance, by far the most likely cause of the bursts' optical dimness is dust inside the host galaxy absorbing light from the afterglow before it escapes, the team concluded. However, the starlight shows no obvious signatures of dust, indicating that the dust may be hiding in patches or clouds where it is difficult to detect.

Consequently, there could be much more dust than has been suspected as the result of measurements using other techniques, and "dark gamma-ray bursts could provide a complementary way of answering the question of how much star formation was going on inside galaxies in the early universe," Perley said.

The authors of the report propose more radio and sub-millimeter observations of the host galaxies of dark gamma-ray bursts to better understand the reasons behind the obscured optical emissions from GRBs.