The universe is in terrible shape -- at least for astronomers grappling with the evolution of the cosmos. Recently, they have watched one of the most promising ideas in modern cosmology -- the so-called "cold-dark-matter" model of how galaxies form -- slowly collapse under its own weight.

"Theorists are beginning to abandon ship," says University of Pennsylvania cosmologist Paul Steinhardt. "And there is no other model ready to step in."

At issue is a deceptively simple question: How did the universe get to be as "lumpy" as it is? That is, how could a cosmos that began some 15 billion years ago in a smoothly expanding Big Bang -- a perfectly uniform cloud of energy and particles -- end up forming the clotted mess of galaxies, clusters of galaxies, and superclusters of clusters that now occupy the heavens?

That question would be much easier to answer were it not for two sets of inescapable and apparently contradictory facts:

The "background radiation" left over from the Big Bang is still observable in the long wavelength microwave portion of the electromagnetic spectrum, and it is exactly the same at every point in the sky to a measured accuracy of about two or three parts in 100,000. That uniformity implies a cosmos in which matter is evenly distributed through space.

Yet the farther scientists are able to peer into the nethermost recesses of the universe, the more "lumps" of large-scale structure they find.

Among the most dramatic recent findings are the "Great Wall" of galaxies found by Harvard astronomers Margaret Geller and John Huchra and the new map of the "local" universe (that is, out to about 500 million light years from Earth) produced by the Infrared Astronomical Satellite.

These and other studies indisputably show vastly uneven clusters of galaxies separated by yawning voids. As a result, says Geller, "we are in this funny situation in which we have two pictures: One of what the universe was like when it produced the microwave background and one of the way galaxies are distributed today. It's like having the beginning of the movie and the end." Search for Plot's Middle

For the past decade, cosmologists have been attempting to fill in the intervening plot -- but without throwing out the Big Bang itself, which remains our most dependable conception of how the universe was formed.

In fact, "the basic Big Bang theory is in the best shape it's ever been in," says David Schramm, a cosmologist at the University of Chicago.

Not only does it accord with the uniform microwave radiation measured by the orbiting Cosmic Background Explorer and other instruments, but it fits the observation, first confirmed by American astronomer Edwin Hubble, that every region of the universe is receding away from every other, and that the farther away a region is, the faster it seems to be receding.

Moreover, the atomic physics at the core of the Big Bang model makes certain predictions about exactly how much hydrogen, helium, lithium and other light elements the primeval fireball should have produced -- which turn out to be precisely the proportions found in the universe today.

That's the good news. The bad news is that the amount of visible matter in the cosmos is not sufficient to account for the way that star clusters and galaxies behave. These objects alone do not contain nearly enough mass -- and hence cannot generate a large enough gravitational field -- to keep themselves from breaking apart.

Some still-unknown entity -- called "dark matter" because it cannot be detected with existing instruments -- must be holding things together. Indeed, at least nine-tenths of the total contents of the universe is in the form of this "missing mass," whatever it is.

Universal Inconsistency

Neither the Big Bang nor missing mass theories, however, can explain the disconcerting lumpiness of the universe. For the past decade, the favorite solution has been "cold dark matter."

Physicists generally agree that immediately after the Big Bang, the seething soup of gases, or plasma, was so hot, and so intensely bombarded with energy, that particles could not hold together long enough to form atomic nuclei; that could only happen after the fireball expanded and cooled. At that point, the universe suddenly became transparent: The light particles, or photons, "decoupled" themselves from the heavy particles and were suddenly free to travel in the intervening space. We detect them now as the "cosmic background."

But according to the cold dark matter theory, weakly interacting particles -- the cold dark matter -- were impervious to radiation, though not to tiny random fluctuations in gravity. Those perturbations caused the cold dark matter to pucker and stretch -- much as scum floating on a pond will bunch together in places when ripples move across the water -- producing slight variations in density.

"Once the photons decoupled," says astronomer Alan Dressler of the Carnegie Institution in Pasadena, Calif., the ordinary particles of matter "were now free to flow into those little pockets that had formed in the dark matter." In time, the theory says, these pockets evolved into ever larger clusters such as galaxies.

This theory has two considerable advantages: It allows for the uniform background radiation and also permits clumping. But it cannot account for the titanic structures astronomers now detect -- at least not in the relatively short time since the Big Bang.

This problem, known for years, did not seriously threaten the theory until earlier this month, when a number of cold dark matter partisans, writing in the journal Nature, reported that their studies of density in the universe showed "more structure on large scales than is predicted by the standard cold dark matter theory of galaxy formation."

'Hot Dark Matter' Theory

This may not be curtains for the theory; tinkering with the assumptions might save it. Dressler, for example, believes that it can't be dismissed conclusively until calculations of the cosmic background radiation are refined further -- a result expected within a couple of years.

Other experts are less sanguine. But "that's okay," says physicist Virginia Trimble of the University of California at Irvine. "There are a dozen other solutions around."

One that has returned to astrophysical fashion lately is the "hot dark matter" model: a primordial plasma populated with a certain kind of neutrino heavy enough to provide the missing mass. But it has a few problems, too. Schramm calls it the kind of theory that "needs two tooth fairies."

However, says Dressler, "Science moves forward by falsifying previous models. . . . It's an incredibly healthy thing when theories are being eliminated. That's how you get the really bright people thinking."