The universe had entered what astronomers characterize as its dark age. Over a period from 300,000 years to about half a billion years after the Big Bang the darkness persisted, denying to astronomers any direct evidence of this time of momentous transformation. Somehow this epoch gave birth to the architecture of the modern universe and conditions that would eventually lead to planets and life.

But what happened in the dark age and how the light came on again is an enduring and frustrating mystery. If there was no light or other detectable radiation from the dark age, what was there to see? Besides, no telescope could see back that far. Sir Martin Rees, a cosmologist at Cambridge University in England, calls this the "most important unsolved problem in astronomy."

In spite of the obstacles, by looking deeper and deeper into space and thus into time with new telescopes in Earth orbit and on the ground, astronomers are making progress in their search for the first light. They are at last catching sight of some of the earlier, if not necessarily the first, galaxies -- each observation seems to put their formation back earlier in time. Astronomers are also gathering clues to what makes up most of the universe, the mysterious dark matter, and how it and other dark-age conditions gave birth to the first stars and the first galaxies. The new knowledge is enabling scientists to test their theories and conduct increasingly detailed computer simulations of cosmic history, including the dark age.

The dark age ended when light returned to the firmament with the emergence of the first stars and then vast communities of stars gravitationally bound in galaxies. "Suddenly, the universe lit up like a Christmas tree," said Dr. Abraham Loeb, a theoretical astrophysicist at the Harvard-Smithsonian Center for Astrophysics in Cambridge, Mass. "Somewhere in the dark age, the transition from the universe's initial simple state to its current complex state occurred. It would be philosophically exciting to glimpse that moment."

A few years ago, Loeb recalled, he was one of a handful of scientists studying the dark age and early star formation. But a workshop on the subject at the University of Maryland last week drew several dozen of the top observational and theoretical astronomers. "Now this has become a very hot topic," he said.

Two weeks ago, the National Aeronautics and Space Administration released pictures taken by the Hubble Space Telescope of what astronomers think are the most distant galaxies ever studied. These galaxies probably formed in the first billion years of cosmic history, or more than 12 billion years ago. Such estimates are rough because they depend on cosmological models and chronological yardsticks based on the universe's being about 13 billion years old, which is controversial and subject to revision.

"We have really made a step into new territory," said Dr. Rodger I. Thompson, a University of Arizona astronomer who directed the observations with the Earth-orbiting Hubble telescope's near-infrared camera and spectrometer.

Only an infrared camera can detect the faintest, most distant galaxies. Dust in deep space obscures most visible light, more so than radiation in the longer infrared wavelengths. And it is a phenomenon of the expanding universe that the light emitted by a galaxy gets stretched out to the redder and even near-infrared wavelengths. At distances of vanishing visibility, the infrared galactic light becomes the cosmic version of the Cheshire cat's grin.

The more distant galaxies have a higher expansion velocity and thus a higher red shift. Since the age of the universe is still uncertain, astronomers express cosmic distance, and thus age, in terms of an object's red shift.

Not so long ago, it was all astronomers could do to see objects at distances at a red shift of 1, encompassing vistas extending the recent half of the age of the universe. When they picked up some red shift-3 sightings, it was the equivalent of aviation's breaking the sound barrier. In recent months, astronomers have broken the red shift-5 barrier, when the universe was only one-tenth its present age.

"We are seeing things at much higher red shifts than we once thought possible," said Dr. John C. Mather, an astronomer at the Goddard Space Flight Center in Greenbelt, Md.

Focused on a tiny patch of sky, the Hubble's infrared camera made a 36-hour exposure and saw more than 100 extremely faint galaxies that were absent in visible-light pictures. At least some of these galaxies are thought to have the highest red shifts ever measured. Precise measurements may have to await the development of more powerful telescopes.

"This is an exciting step into the infrared view of the universe," said Dr. Alan M. Dressler, an astronomer at the Carnegie Observatories in Pasadena, Calif. "That's where we're going to be looking for galaxy formation."

The first big break in distant observations came in 1995, when Dr. Robert Williams, director of the Space Telescope Science Institute in Baltimore, directed astronomers to concentrate the Hubble on a narrow sector of the sky, a region virtually devoid of light as seen from the ground. The telescope's visible-light camera took a 10-day-long exposure, long enough to resolve objects at then-record distances.

The Hubble Deep Field, as the resulting multicolor picture is called, gave astronomers a remarkable view of more than 3,000 galaxies, some at distances beyond red shift 4. They were astonished to find such a dense population of galaxies so far away and so early in time. It meant that astronomers had a way to go to reach the time of galactic genesis, the frontier to the dark age.

At about the same time, a team of astronomers led by Dr. Charles C. Steidel of the California Institute of Technology in Pasadena developed a novel technique for detecting faraway galaxies using the world's most powerful telescopes at Keck Observatory in Hawaii. By now, the astronomers have confirmed the presence of 440 high-red shift galaxies existing just a few billion years after the Big Bang. No objects at such distances had ever been examined in such close detail.

Steidel's observations are made by the ultraviolet "dropout" technique. Ultraviolet light from a remote galaxy tends to be absorbed by hydrogen within the galaxy itself and between there and Earth. The more distant a galaxy, the more hydrogen its light encounters in the intervening space, and thus the more its ultraviolet emissions are dimmed. By seeking objects that show up in green and red filters but not in ultraviolet, the astronomers detected galaxies with red shifts from 2.2 to 4.

The Hubble Deep Field and Dr. Steidel's snapshots have given theorists much to think about. The early galaxies appear to be smaller, more concentrated and less regular than today's galaxies. And they are "as strongly clustered in the early universe as they are today," Steidel said.

Small, irregular galaxies, cosmologists say, fit their models in which structure in the universe evolved from the bottom up, and not from the top down. That is, galaxies presumably started as small collections of stars and these merged to form much larger objects, instead of stars starting out in extraordinarily vast congregations and then fragmenting into galaxies.

"The small, young objects we find are still in the process of formation," Loeb said, explaining why the findings seem to support a growing consensus in favor of the bottom-up interpretation.

The observations also suggest that cosmologists are on the right track with their ideas about the nature and amount of cosmic mass. Years of studying the rotations of galaxies near and far have convinced scientists that there is much more to the universe than what they can see. Galaxies would have disintegrated long ago if they were not held together by more gravity than ordinary matter can account for. Most of the universe's mass, maybe as much as 99 percent, must be in some mysterious "dark" form.

In "Before the Beginning," published this year by Perseus Books, Dr. Rees of Cambridge writes, "Clearly we will never understand galaxies properly until we understand the dominant stuff whose gravity binds them together."

The favored hypothesis predicts the existence of cold dark matter, exotic and as yet unknown subatomic elementary particles that emerged from the Big Bang. They are so small and slow that they virtually never interact with each other or with ordinary matter except through gravity and thus produce no detectable radiations. Yet they could have supplied the decisive gravitational force to marshal ordinary matter into galaxies and clusters of galaxies.

It all began, according to theory, with tiny ripples in the cosmic microwave background radiation, the detectable afterglow of the Big Bang that is the source of much knowledge about the universe's first 300,000 years. These slight fluctuations in radiation, confirmed and plotted by COBE, the Cosmic Background Explorer spacecraft, in the early 1990's, are caused by irregularities in the density of matter in the Big Bang. Such clumps of dark matter would have attracted ordinary matter to create larger clumps and set in motion agglomerations leading to stars and galaxies.

Computer-generated simulations based on the COBE findings and many other physical data and assumptions have showed that a universe dominated by hot dark matter, an alternative idea invoking known fast-moving particles called neutrinos, should have resulted in a top-down creation in which much larger conglomerations of matter would form first and then break up into galaxies. This scenario, however would result in a much longer dark age than now appears to be the case. Cold dark matter, on the other hand, should have formed galaxies on small scales and at a relatively early time. Which is just what the new observations are revealing.

In February, astronomers announced the first detection of a red shift-5 object, a young galaxy that existed when the universe was only 8 percent of its present age. They had crossed the frontier into the first billion years of the universe, with the boundary to the dark age still not in sight.

Working at Keck Observatory, a team led by Dr. Arjun Dey of Johns Hopkins University in Baltimore measured a red shift of 5.34 for the early galaxy, meaning that it was shining 820 million years after the Big Bang (also assuming an age of 13 billion years for the universe).

"Now that we know what to look for, I'm sure this record will be broken in a matter of months," Dey remarked at the time.

Sure enough, in May astronomers at the University of Hawaii and Cambridge University reported sighting a more distant galaxy with a 5.64 red shift, pushing back galactic discovery about 60 million years. One of the discoverers, Dr. Esther M. Hu, said astronomers were on the verge of seeing galaxies back to about 500 million years after the Big Bang.

Loeb noted two intriguing clues that light had returned to the universe at least that early, perhaps at red shifts beyond 10. Both came from observations of light from quasars, cores of distant galaxies and the brightest objects in the universe. Because the quasar light passes through clouds of hydrogen gas before it reaches Earth, it serves as a probe of the cosmic environment out to great distances.

The fact that the quasar light is not absorbed by the gas, Loeb said, is evidence that nearly all of the hydrogen atoms in deep cosmic space have been ionized, their electrons stripped away by ultraviolet radiation. Such radiation could have come only from a large population of stars at red shifts beyond that of the quasars. Other quasar studies yielded the second clue: some of the distant gas clouds contained not only hydrogen but heavy elements like carbon, which are produced only by the nuclear furnaces inside stars and scattered through space when stars explode at the end of their lives. So there had been stars living and dying long before these early galaxies.

"We are going to fainter and fainter objects and getting smarter and smarter each month," said Dr. Dressler of the Carnegie Observatories. "But we've pushed just about as far as we can expect from today's telescopes, even the Keck or the Hubble."

Astronomers are in the early stages of planning a more powerful replacement for the orbiting Hubble telescope. Called the Next Generation Space Telescope, it is expected to be launched a million miles into space in 2007 and be able to detect infrared emissions from objects 400 times fainter than those being studied with any of today's ground-based and orbiting infrared telescopes. Its mirror will be nearly three times larger than Hubble's.

With the new telescope thus able to gather so much more light, Mather of the Goddard space center said, astronomers in exposure times as short as two weeks should "see as far as you can imagine seeing." They should be able to make out galaxies with red shifts of 10 or more.

That might be just enough to observe the first stars and galaxies and understand more about how the early universe passed its long night of darkness organizing itself to bring form and light everywhere.