- "Fire and Ice," Robert Frost, 1923

Ice wins. After years of observations using telescopes all over the globe, and more recently the Hubble Space Telescope, the international Supernova Cosmology Project led by astrophysicist Saul Perlmutter has gathered enough compelling evidence to predict the fate of the universe.

n 1917, Albert Einstein and other physicists believed our own Galaxy was all there was to the universe-a uniformly dense collection of stars and other matter floating in the void. The problem was, Einstein's new General Theory of Relativity would not allow for a static universe.

So Einstein fudged; to get rid of the possibility of expansion or contraction, he introduced into his equations a term called the cosmological constant. Although other solutions were also possible, the value he gave the cosmological term at least permitted the universe to be static.

Edwin Hubble revealed in 1929 that the universe was full of numberless galaxies, that their light was redshifted-its frequency reduced, like the lowered pitch of a fire truck's siren as it races away-and that the light from more distant galaxies was redshifted farther than closer ones. It was immediately clear that the universe really is expanding.

Einstein called the cosmological constant "the biggest blunder of my life." Doubly so: not only was it unneeded, without it he might have predicted the expansion of the universe.

The question is not as simple as Edwin Hubble's first observations suggested, however. The redshifts of distant objects are greater than their distances imply if the universe were expanding at a constant rate. In fact the expansion rate was higher in the past. But how fast is it slowing down, and is the deceleration continuing?

(Click on image to enlarge)

Two images from the Cerro Tololo Observatory in Chile show a region of sky just before and after a supernova appears. The third image shows the same supernova as seen by the Hubble Space Telescope, a sharper picture that allows a better measurement of the supernova's apparent brightness, and thus its distance.

If the universe is sufficiently dense, someday the galaxies will stop their outward rush and fall back upon themselves in a Big Crunch. Less density, and the galaxies will escape to infinity, resulting in a Big Chill. It's also possible that the galaxies will slow ever more gradually, almost-but never quite-coming to rest. This is the alternative favored in the popular Inflationary theory proposed in 1979 by Alan Guth, which nicely cleans up a number of messy questions about the structure of the universe and its parts.

The Supernova Cosmology Project looks into the distant past by looking at distant objects in space. To estimate distance, astronomers use "standard candles," objects whose intrinsic brightness is the same wherever they are found (named to suggest an ordinary wax candle that, if it is seen across a room, for example, looks dimmer than it does up close). Type Ia supernovas are superb candles, stellar explosions so bright that for a few days they can be brighter than entire galaxies. They can be seen across billions of light-years.

First astronomers have to catch the supernovas in the act of exploding-a rare chance event until the Supernova Cosmology Project developed a novel technique. Ground-based telescopes in Chile and elsewhere take pictures of small patches of sky, enough to include a thousand galaxies at a time. The same fields are recorded again three weeks later, both to take advantage of the dark of the moon, and, says Gerson Goldhaber, who helped develop the technique, "because we want to catch the supernova before it reaches its maximum brightness." Any new bright spot in the field is a supernova candidate.

"Prior to 1992, before the technique was fully evolved, no very high-redshift supernovas had been found," says Goldhaber, but with the new method "and better access to powerful telescopes, the rate of discoveries has been increasing exponentially." Team members using the Keck telescope in Hawaii over the 1997 Christmas holidays discovered 15 new Type Ias, bringing the total so far to more than 60.

Ground-based telescopes peering through the Earth's shifting atmosphere can separate the spectrum and brightness of an extremely distant supernova from that of its home galaxy, but only with difficulty. "We first proposed using the exquisite imaging capabilities of the Hubble Space Telescope two years ago," says Saul Perlmutter, "once we had proven that we could supply it with large numbers of distant supernova discoveries 'on demand.'"

Lead scientists on this project

Learn more about these researchers

On January 8, 1998, Perlmutter announced the results of the first 40 supernovas the team had analyzed, including one studied with aid of the Hubble Space Telescope. The brightness of the most distant revealed that its light had been traveling toward Earth for seven billion years, half the age of the universe. Together, their redshifts showed that the universe had been expanding faster in the past, but it wasn't slowing down fast enough to stop-and in fact may now be starting to accelerate again.

"All the indications from our observations of supernovas spanning a large range of distances are that we live in a universe that will expand forever," Perlmutter says. "Apparently there isn't enough mass in the universe for its gravity to slow the expansion, which started with the Big Bang, to a halt."

But if Inflationary theory is correct (perhaps even if it isn't), some factor besides gravity may be involved.

"With a large enough sample of data of the kind we can expect routinely from the Hubble Space Telescope," Goldhaber says, "we now believe it will be possible to independently measure the effects of gravity and the cosmological constant."

When Einstein called the cosmological constant his biggest blunder, he may have spoken too soon.

- Paul Preuss