The data also are giving cosmologists greater confidence in conclusions about the structure of the universe gleaned from studies of distant supernovas. These include recent findings that the universe is flat, its expansion is accelerating, and that some two-thirds of the universe's mass is made up of strange matter called dark energy or "vacuum energy."

"Our study of distant supernovas depends entirely on a reliable calibration of nearby ones," said Alex Filippenko, a professor of astronomy at UC Berkeley and leader of the Lick Observatory Supernova Search (LOSS) conducted by the Katzman Automatic Imaging Telescope (KAIT). "We want to understand the basic physics of supernovas going off in galaxies, and KAIT is so important because it finds essentially all the supernovas in the galaxies we monitor."

Filippenko summarized the operation of the robotic telescope and results to date at a media briefing Monday, June 5, during the national meeting of the American Astronomical Society in Rochester, N.Y. His formal talk at the meeting takes place at 2 p.m. on Monday in a session on automated telescopes. More details will be presented by recent UC Berkeley graduate Maryam Modjaz in a display paper the same day.

UC Berkeley's robotic telescope scans about 1,000 galaxies a night looking for new points of light, generally within about 500 million light years (3 billion trillion miles) of Earth. Put into operation in late 1996, it was fully programmed to conduct the supernova search by the middle of 1998. By the end of 1998 the KAIT had found 19 new supernovas, and in 1999 it found 40 - nearly three times the number found in a calendar year by any previous robotic supernova search. This year it already has discovered 11 new nearby supernovas.

Because the telescope catches almost every supernova in the galaxies it monitors, the data give astronomers a better idea of how common supernovas are and of their variety.

Filippenko and post-doctoral fellow Weidong Li have found, for example, that so-called Type Ia supernovas are less uniform than once thought. This may have major implications, since Type Ia supernovas are used as a "standard candle" to estimate galactic distances. If there are several common types, conclusions based on an assumption of uniformity may be incorrect.

Type Ia supernovas are thought to occur in binary star systems. One of the stars, a white dwarf, is thought to steal matter from a larger companion until its mass exceeds a certain limit, at which point the star becomes unstable and is consumed in a gigantic thermonuclear explosion.

"This is the process by which we are made of star stuff," Filippenko said. "Iron and oxygen and other heavier elements are produced in supernovas. This is where we came from."

Based on their supernova data, he and Li have concluded that more than one-third of Type Ia supernovas are peculiar, in that they are brighter or dimmer than the "average" Type Ia, or that their spectra show unusual chemical composition.

They and their colleagues have submitted a paper to the Astrophysical Journal documenting the high percentage, about 36 percent, of peculiar supernovas, and speculating on what this implies about what is creating the explosion. It may be, for example, that some Type Ia supernovas result from the merger of two white dwarf stars into an unstable mass that quickly explodes. Thus, Filippenko said, Type Ia supernovas may be produced in multiple ways.

Filippenko and his colleagues also have found possible differences between the pattern of brightening of nearby supernovas and the pattern typical of distant supernovas. He doubts this is a significant problem for cosmologists who have concluded, from analysis of the distance and recession speed of distant supernovas, that the universal expansion is accelerating. Nevertheless, he said, it is important to verify that supernovas which exploded billions of years ago look the same - in particular, that they have the same peak power output - as those which exploded more recently.

"When we say some supernovas are further away then we expect, we are making an assumption that they aren't dimmer than we expect," said Filippenko. "Confidence in our conclusions can only get better with a fair comparison of distant and nearby supernovas."

Filippenko is part of an international crew of astronomers - the High-Z Supernova Search Team - that in 1998 concluded from a study of Type Ia supernovas that the expansion of the universe is accelerating. That information, combined with new data on the cosmic background radiation, recently led cosmologists to even more amazing findings about the universe: not only is it accelerating, but its geometry is flat. For this to be the case, the normal matter we see and the dark matter we infer to be present from other studies comprise only one-third of all the matter in the universe.

"Our earlier work on supernovas showed some weird stuff in the universe, essentially a 'vacuum energy' that is accelerating the expansion," he said. "Now the MAXIMA and BOOMERANG cosmic microwave background radiation experiments have given us high quality data that complement the supernova data, and create a greater acceptance of this weird stuff in the universe. This will create a revolution in the next decade."

To boost confidence in their analysis of distant supernovas, Filippenko and his team have secured time this fall on the Hubble Space Telescope to obtain accurate color measurements of seven distant supernovas. They hope that by comparing the colors of distant and nearby supernovas, they can determine if there is any intervening dust that would dim the light from distant supernovas and make them appear farther away, thus throwing their conclusions into question.

Filippenko and Li also have embarked on a study of the motion of nearby supernovas to get an idea of the bulk flow of galaxies and clusters of galaxies through the universe. This will provide a measure of the amount of normal and dark matter in the universe, Filippenko said.

The recent successes of KAIT are largely due to Li, placed in charge of telescope operations in 1997 after arriving from the Beijing Astronomical Observatory, where he had directed their automated supernova search. UC Berkeley engineer and astronomer Richard Treffers made significant additional improvements in the hardware.

Every night computers check the weather at Lick Observatory on Mt. Hamilton outside San Jose, Calif., open the dome, point and focus the telescope and take digital pictures. The images are sent via the internet to Li's computer at UC Berkeley, where they are compared automatically to earlier pictures of the same region of sky. Images with new points of light are flagged so that, the next morning, a crew of several undergraduates can double-check them and identify the best candidates. In winter, the KAIT can take as many as 1,200 images a night. Li is working to increase this number, and so far has reduced the time per image from 40 seconds to 30 seconds.

When a supernova is discovered, spectra of the star often are obtained at the Shane three-meter telescope at Lick Observatory, and photometry is carried out by KAIT itself.

KAIT construction and operations have been supported by the Sylvia and Jim Katzman Foundation, which donated funds at a critical time in the telescope's development, as well as by the National Science Foundation, the National Aeronautics and Space Administration, the University of California, the Hewlett-Packard Company, Sun Microsystems, Photometrics Ltd., AutoScope Corporation, and Lick Observatory.

NOTE TO EDITORS: Digitized photographs may be downloaded for use in news stories at the URL http://astron.berkeley.edu/~bait/kait.html.

Media Contact: Robert Sanders, UC Berkeley Public Affairs, 510-643-6998; rls@pa.urel.berkeley.edu

Alex Filippenko, on June 5 and June 6 Filippenko can be reached in Rochester at 716-546-6777; after June 6, he can be reached at 510-642-1813; alex@astro.berkeley.edu

Weidong Li, 510-643-8973; wli@astro.berkeley.edu