SIMULATIONS SUGGEST TEMPERATURE FLUCTUATIONS SHAPED EARLY UNIVERSE.

Byline: John Noble Wilford The New York Times

Even their most successful theories and the stunning discoveries by new telescopes on the ground and in space still leave cosmologists with fundamental gaps in the history of the universe they seek to reconstruct. Hard facts are scarce and subject to conflicting interpretations. Cosmologists are left perplexed but undaunted.

One of the most revealing ways - sometimes the only way - they have of compensating for insufficient evidence and testing alternative ideas about the evolution of the universe is through simulations using the most powerful supercomputers. They feed into the computers certain observed and assumed conditions in the early universe and see what this could lead to over billions of years.

In what is described as the most ambitious such computer simulation, an international group of astrophysicists led by Carlos S. Frenk of Durham University in England has demonstrated how observed temperature fluctuations in the primeval universe could lead from an almost uniform fireball to today's cosmic panorama of galaxies and awesome clusters of galaxies among vast cosmic voids.

The simulations seemed to support a growing realization among astrophysicists that the mass density of the universe may be less than they had assumed for years. Also, they fostered a revived interest in an idea that Einstein proposed more than 75 years ago and then discarded as ``the greatest blunder of my life.'' But the research, scientists emphasized, did not undercut any important tenets of the big bang, the reigning theory of how the universe began in an explosive fireball 10 billion to 20 billion years ago.

Frenk is to report the simulation findings on Friday at a conference of British astronomers in Liverpool. In an announcement Monday, the Royal Astronomical Society called the simulation ``one of the most significant ever undertaken for our understanding of the natural world.''

In interviews by telephone last week, Frenk conceded that the research would not fully answer any of the nagging questions about the expansion rate and age of the universe, the fate of the universe or the nature and density of the pervasive ``dark matter'' that is far more abundant than visible ordinary matter. But, he said, the simulations pointed out strengths and weaknesses in several theoretical models of cosmic evolution.

``Our simulations give us greater confidence in what are, you might say, the best-buy models of the universe,'' Frenk said. ``Some of the tested models produce gross morphological features that are very close to what we see in the real universe. Others simply do not.''

The starting point for all the models was the universe as it presumably existed 300,000 years after the big bang. As predicted by theory, the initial explosion left an afterglow of radiation throughout the sky, the microwave background radiation, which was detected in 1965.

As further predicted by theory, the cooling, expanding universe was imprinted with tiny energy fluctuations after 300,000 years, wrinkles in an otherwise smooth cosmic fabric. These were detected by an American satellite in 1991 and immediately interpreted as the fossil record of the seeds from which galaxies formed.

Beyond that, different assumptions were introduced into four models, particularly different estimates for the density of the universe.

This tested an idea, favored for the last 15 years, that not only is the universe made of a substantial amount of invisible and presumably exotic dark matter, but that there is enough of it to give the universe a critical density - just the right amount of mass to keep the universe from collapsing of its own weight or expanding into infinity.