Why is the universe filled with the light of shining stars and not black holes, or small, dull spheres? Mysteries surrounding the birth of stars may be illuminated by a California astronomer's new theory.

The theory is the first to connect a series of puzzling discoveries made by numerous scientists over the last decade, some in the last year or so. It places these discoveries in a certain order and timing, to depict a four-stage, 10 million-year drama that begins with cold dust in space and peaks when intense stellar winds blast away veils of gas to reveal a burning star.

The most controversial aspect of the scenario is its explanation of why most young stars are much smaller in mass -- by thousands of times -- than the gas clouds from which they were formed.

The theory, developed by Frank Shu, chairman of astronomy at the University of California at Berkeley, holds that a young star limits its mass by becoming a kind of pressure cooker -- like a pot full of boiling water with the lid on. The pressure ejects most of the mass as powerful twin jets of gas shooting from the star's poles -- the vents in the "lid" -- back out into space at 500,000 mph.

Thus most of the billions of stars are of "modest" size, similar to our sun, and provide the conditions necessary to produce the thermonuclear fusion that makes possible the universe as we know it. (Star sizes vary much less than their brightness.)

If there were no mechanism to limit the star sizes, they would become so huge that, most likely, they would "all become black holes, and there would be no light in the universe," Shu said in an interview. A black hole is an ultimate state of gravitational collapse that occurs in huge stars, whose matter becomes so dense that not even light can escape its pull.

If stars were much smaller, they would "stay cold, like the planets, and never shine. There would be no energy source in the universe," he said.

Shu's theory, in a sense, credits nature with knowing how to set up the right conditions for thermonuclear fusion. "Otherwise you have the conclusion that the origin of stellar masses is completely a coincidence," he said. Since wherever one looks in the universe, stars are the primary constituents, he added, "that woud be quite a coincidence."

Enveloping gas clouds have hidden most of this process from human probing. But in the past 15 years, aided by new techniques using rockets and satellites, scientists making infrared and radiotelescope observations have looked through the clouds and caught pieces of the action that provide important new clues. For instance, they've seen the jets of gas but not their sources.

Scientists have found that before baby stars light up their thermonuclear fires and begin to shine, they are quite energetic, with great stellar winds, magnetic activity and flares some 10 million times greater than more mature stars such as our sun.

But how these forces interact to create a burning star and the order in which they occur have never been positively established.

Philip Myers, of the Harvard-Smithsonian Center for Astrophysics in Cambridge, Mass., said, "What makes Frank's work different is that he has a comprehensive picture of the four major stages of the birth process . . . . It's a much better integrated theory than had previously been suggested."

The starbirth drama begins with the formation of small, dense pockets of cold dust and gas that slowly condense out of vastly larger clouds. This stage, according to Shu and others, lasts 1 million to 10 million years, which explains why stars in groups show this spread of ages.

As each of these pockets of matter grows, its gravity causes it to collapse from the inside out -- "the dense region near the center collapses first and outer layers begin to fall inward in waves."

For perhaps another 100,000 years, a "protostar" at the center grows as its gravity draws in more gas and dust.

Until the last year or so, the evidence of this "infall" was too small to be detected by radioastronomers. But beginning in 1985, using a new generation of radiotelescopes, astronomers in Massachusetts and Arizona observed a pattern that, Shu said, "does indeed support the notion of an inside-out collapse."

Much of the incoming gas and dust fall into a flat disk swirling around the forming star at the center, following the motion of the original cloud core, Shu said, and the whole mixture heats up.

The kind of infrared radiation that ought to emerge at this point, if Shu's scenario is correct, was found in observations of interstellar clouds by the Infrared Astronomical Satellite (IRAS), launched in 1983.

Until then, "there was no good set of theoretical predictions that matched the emissions from the youngest stars that we know about," Myers said. While it is not conclusive proof, he added, "it makes a much stronger case than we had before."

The theory's most dramatic -- and so far least documented -- aspect is its explanation of how the inflow of gas that feeds the star is reversed to form the jets that stabilize the star's size.

As the emerging star expands and grows hotter, deuterium -- a form of hydrogen near its center -- begins a fusion reaction that generates powerful blasts of hot gas and other activity moving outward. The upwelling forces run into the "lid" of inflowing cloud material "like water boiling in a pressure cooker" until jets of gas break through at the lid's weakest points -- the star's poles.

The star would still be obscured in the cloud of gas during this phase, Shu calculates, with only the bi-polar gushers of gas -- the puzzling jets observed recently by astronomers -- visible at interstellar distances.

The lid of inflowing gas weakens as more and more of it falls into the whirling disk around the forming star. Gradually the forces of the gushers of gas shooting out and the inward-pushing lid-cloud reach an equilibrium. The jets begin to spread out and eventually force their way through the gas cloud in all directions, "sweeping the shroud away and revealing the star and its surrounding disk."

Evidence for this last phase, which Shu said may last from tens of thousands to 100,000 years, comes from observations of fledgling stars, from a pattern in the ratio of their mass to their radius.

"Stars then -- and not their parent clouds -- determine their own ultimate mass," Shu concluded in the paper on his theory, which he presented to the International Astronomical Union meeting in West Germany.

"It is extremely fortunate that nature automatically makes objects with the mass scale capable of thermonuclear fusion," he said. "Otherwise, the universe would include no chemical elements other than hydrogen and helium" -- the stuff of interstellar gas -- and there would be no stars to carry out the fusion process that creates all the heavier elements, and the molecular ingredients of life would not exist.