Princeton University astronomers have found evidence that a mysterious, extremely large mass is drifting in space halfway between Earth and the edge of the observable universe.

Although the object has not been seen directly, the discoverers estimate that its gravity is that of a thousand large galaxies combined.

The discovery near the constellation Virgo could lead to a revolution in understanding of the universe. The object is impossible to classify, according to Edwin L. Turner, a Princeton astrophysicist and principal author of a report on the find in this week's issue of the British journal Nature.

"The most conservative explanation would be that it's a cluster of galaxies as large and dense as any we've ever seen," he said. "The problem is that we should be able to see such a big cluster quite easily in our telescopes, and we don't." Observations with more sensitive radio telescopes have also failed to see evidence of such a cluster.

The other possibilities, considered extremely exotic by astronomers, are a black hole with the mass of 1 quadrillion stars and something called a cosmic string, a sort of crack in space itself.

"We may be seeing something that has only existed in our theories until now," said J.R. Gott, another Princeton astrophysicist.

Evidence of the new object, gathered in early March with the four-meter Mayall optical telescope at Kitt Peak National Observatory near Tucson, is in the form of a double image in the sky. What seems to be two quasars, galactic cores so bright they are visible even at the universe's edge, is really just one.

The double image, a kind of space mirage, is caused by an intervening object whose gravity is so powerful that it bends the quasar's light, like a distorting lens. Predicted by Albert Einstein and others in the 1930s, the phenomenon was found first in 1979 and has been seen five times since. None of these so-called gravitational lenses, however, has been as powerful as this one.

The newly discovered lens separates the quasar images by about 2 1/2 minutes of arc, about 30 times the separation caused by the other lenses. The full moon is about 30 minutes of arc, or half a degree, in diameter.

Astronomers have said they are convinced by the size of the separation that the unseen object, the source of the gravity causing the double image, must be unprecedentedly massive. When the dual image was first noted in 1979, it was not investigated as a candidate for the lens effect because of the improbably wide separation.

Turner acknowledged a slim possibility that the double images might be two distinct objects but said the evidence makes this extremely unlikely. Spectral analyses of both images, in which their light is separated into its component colors, show that each has essentially identical component wave lengths. Quasars usually are quite different in spectra.

"It would be too much of a coincidence for two quasars to be so nearly identical," he said.

Jeremiah Ostriker, chairman of Princeton's department of astrophysical sciences, agreed, saying, "I don't think nature is playing tricks on us."

"Whatever the object is, its mass makes it quite unusual," said Alexander Vilenkin, a Tufts University physicist. "If this actually is a lens, its enormous separation means we will almost certainly learn a great deal about the way the universe has evolved."

Word about the new lens has spawned a flurry of papers exploring its consequences for prevailing theories about the universe. The 200-inch telescope atop Mount Palomar in California is being trained toward the lensing object and is also seeking evidence that other, fainter nearby quasars may have been lensed into double or triple images.

The number of images is important. If the lens is a cluster of galaxies, it will produce three or more images. If it is a black hole or cosmic string, it will produce two.

Moreover, all pairs of string-generated images would be arrayed with one image on each side of an invisible line. But if the object is a black hole, the images would be paired around a common center. Preliminary searches have revealed only the pair of images, making it impossible to answer the question.

Cosmic strings and black holes are two of the newest additions to astronomy's menagerie of theoretical objects in the cosmos. Both are thought possible, but neither has been seen for certain.

Of the two, black holes are thought more likely. They are born when stars run out of fuel and collapse. If the star has enough mass, collapse is inexorable. Instead of stopping when the mass is compressed into a solid, compression continues indefinitely, beyond the point at which atoms lose their identity.

Finally, the gravity is so strong that light cannot escape. The star, now only a few miles wide, is black. Black holes can grow to enormous size by pulling in other stars, but astronomers calculate that it would take longer than the universe has existed for one to accumulate enough mass to be as powerful as the newly discovered object. Cosmic strings, much more speculative, are strings of matter left after formation of the universe and far thinner than a subatomic particle but with a mass of billions of tons per inch of length. "Cosmic strings are a very exciting possibility in this case," Gott said, "because they should produce images with just this kind of [image] splitting."

Astronomers at AT&T Bell Labs in Holmdel, N.J., are seeking evidence of a cosmic string by searching for distortions that it would theoretically produce in faint microwave radiation that pervades the universe as a leftover from the "Big Bang" explosion in which the universe is believed to have begun. The radiation's intensity from one side of the string to the other should change slightly.

"Our detector is more sensitive than any other currently set up," said Bell Labs' Anthony A. Stark, "and we don't see any compelling evidence either way." Stark and Robert W. Wilson, who won a 1978 Nobel Prize for discovering this background radiation, plan more detailed observations next year.

The background of microwaves also can give clues about the other candidates. If the lensing mass is a cluster of galaxies, its gas should produce a noticeable dimming of the microwaves. But Bell Labs' search has not found this.

If the mass is a black hole, it will appear as a void in the sky but too small to be seen with anything but the Very Large Array radio telescope near Socorro, N.M., which may be used this summer.

If none of the proposed theories about the mass proves correct, the unsolved puzzle could keep astronomers busy for years.

"My personal guess," Princeton astrophysicist Bohdan Paczynski said, "is that this is what will happen."