ith test after test currently showing complete agreement with Einstein 's general theory of relativity, it's natural to ask why physicists put such effort into designing ever more sophisticated experiments—because, says general relativist Clifford Will, “every test of the theory is potentially a deadly test.” Gravity is the force that rules the universe. To understand its workings, to the finest degree, is to understand the very nature of our celestial home. Any deviation, any surprising signal, would surely offer clues to an even deeper understanding of the cosmic design. Consequently, gravity wave astronomers will not be satisfied with earth-based interferometers alone. Interferometers on the ground are limited in the range of frequencies they can detect, which is unfortunate since many interesting events originating from strong astrophysical sources are expected to occur in the very low frequencies, from a millionth of a hertz to 1 hertz. To examine those regions, gravity wave astronomers must venture into space.

In some ways, gravity wave astronomers have already established a presence in space with the many spacecraft that have or at this moment are speeding through the solar system on their journeys to various planets. In each case there are two masses: one is the spacecraft; the other is the Earth. From Earth a signal with a very precise frequency is transmitted to the spacecraft, and a transponder aboard the spacecraft sends it back. If a gravity wave passes by and jiggles the Earth, it will shift the frequency of the signal it is sending out ever so slightly. Whenever the gravity wave hits the spacecraft, it will cause a similar frequency change in the spacecraft's returning signal. All in all (as soon as all other sources of noise are subtracted out), what's left is a distinctive set of pulses, the imprint of the radio wave being intermittently altered by the gravity wave. The entire Earth/spacecraft system acts in some way like an interferometer with a single arm hundreds of millions of miles in length.

Planetary scientists have already monitored the communications of the Viking, Voyager, Pioneer 10 and 11, Ulysses, and Galileo spacecraft. The chances were always small that anything would be detected; fluctuations in the solar wind can alter the frequency of a radio beam. But if a particularly big gravity wave had passed by during those missions, the opportunity was there. The gravity waves capable of being found by this method would be extremely long. From peak to peak one full wave could stretch from here to the Sun or even farther. Such a lengthy undulation might have been sent out by two supermassive black holes in close orbit around one another in some far-off galaxy.

One of the biggest tests of this kind was carried out in the spring of 1993. At that time three separate spacecraft were speeding from Earth in different directions, which offered the perfect opportunity to test for gravity waves. NASA's Mars Observer (before it mysteriously malfunctioned) was headed toward the red planet, the Galileo probe was trekking toward Jupiter, and the European Space Agency's Ulysses was journeying toward the Sun. Between March 21 and April 11 of that year, radio signals were simultaneously beamed toward the three spacecraft using a network of radio antennas situated around the globe (NASA 's Deep Space Network). Once received, each craft amplified the signal and sent it back to Earth. The American and Italian plan-

A gravity wave arrives from a cosmic event. One way to detect the wave is to compare how the wave affects the speed of a clock on Earth, then one on a spacecraft situated in the far solar system.

etary scientists running the test figured that if a gravity wave had passed by the spacecraft would have gently rocked, like buoys bobbing on the ocean of space-time. That is why the frequency of the radio transmission would have shifted, ever so minutely. A change from just one spacecraft might have been spurious, the result of a local disturbance. But if a frequency shift had been detected in the signals from all three probes around the same time, the evidence would be more convincing. Superaccurate atomic clocks, timepieces capable of discerning a change in frequency of a few parts in a million billion, were used to monitor the potential shifts. Nothing resembling a gravity wave was seen, but that was not too surprising. It was a long shot at best. Spacecraft are still buffeted by the solar wind as they cruise through space, and the Earth's turbulent atmosphere can introduce some radio noise as well.

For the 1993 test the frequencies used to communicate with the spacecraft were between 2.3 and 8.4 gigahertz. But the use of an even

higher frequency should help cut down on interference from the plasma in the Earth's ionosphere as well as the solar wind. That will be the case with the Cassini mission, a spacecraft launched in October 1997 and now on its way to study Jupiter and Saturn. “We were hitchhikers before,” says John Armstrong of NASA's Jet Propulsion Laboratory (JPL). “But now on Cassini, there is specific hardware on board to carry out these gravity wave experiments.” Cassini will use a frequency of 32 gigahertz, four times higher than in previous space tests. There will be three opportunities to search for gravity waves between Cassini and Earth: 40-day periods from December to January in 2001, 2002, and 2003. When such tests were first done in the 1980s with Pioneer 10 and 11, the potential gravity wave strains that the Earth/spacecraft system could measure were around 10^-13. Cassini has the capability to do a thousand times better, down to a strain of about 10^-16, largely due to its use of the higher communication frequency. That 's also 10 to 30 times better than the 1993 experiment. The Cassini experiment, in fact, will be about the best that can be done with this method of testing. To do better will require sending an entire interferometer into space.

In the fall of 1974 that topic came up during a meeting that Weiss convened for his NASA committee looking into gravitational space physics. A group from NASA's research facility in Huntsville, Alabama, made a specific proposal to send a huge laser interferometer into space. Two aluminum trusses, each about a kilometer in length, would be manufactured in space and then put together in the form of a cross. The test masses would be suspended from this structure. It was a time when scientists were eagerly discussing the prospects of sending a variety of instruments into space. After the meeting, over dinner at a Boston seafood restaurant, Weiss asked Peter Bender whether such a gravity wave antenna sounded like a sensible idea. During the ensuing conversation, they began to think of using separate spacecraft rather than a rigid structure, which would allow the mirrors to be many kilometers apart.

Bender was a member of Weiss's committee because of his previous work on a lunar ranging experiment. Trained as an atomic physicist by Dicke at Princeton in the 1950s, Bender went on to work at the National Bureau of Standards and later with the Joint Institute for

Laboratory Astrophysics at the University of Colorado in Boulder. His work on precision distance measurements commenced in 1962, when he began a long-term collaboration with James Faller, another student of Dicke's who had newly arrived at JILA as a postdoc. Faller was eager to convince NASA to deposit a package of reflectors on the surface of the Moon during an unmanned lander mission in order to reflect a laser beam sent from Earth. By measuring the travel time as the beam bounced between the Earth and the Moon, researchers could determine the lunar orbit very accurately, as well as learn much about the lunar interior. Later it came to be seen that the mirrors could be used to carry out tests of general relativity as well, particularly whether the Moon and the Earth differed in their acceleration toward the Sun. By 1965 Faller's idea, with others brought on, turned into the Lunar Ranging Experiment (LURE).

Sheer luck allowed the LURE team to piggyback its project onto the Apollo 11 mission, the first manned landing on the Moon. Worried that the Apollo 11 astronauts would not have time to carry out all of their planned experiments, NASA began looking for projects that didn't require much setup time. The LURE proposal was perfect. All the astronauts had to do was set the reflector package on the lunar surface and adjust the mirrors to the proper angle to reflect a laser beam from Earth. LURE has been a very hardy experiment. Indeed, the system is still working today, the last experiment to continue operating from the Apollo program. The passive mirrors just keep on reflecting, as there is as yet no evidence of damage from dust or micrometeorites. Laser tests with it continue to be carried out from a French observatory near Grasse and from McDonald Observatory in Texas. “It's tested Einstein 's Strong Equivalence Principle to a part in a thousand. If you look at tests of relativity, that's one of the major ones that's been done,” notes Bender. It shows that gravity accelerates objects equally, regardless of their mass or energy. The Earth and the Moon have been found to accelerate toward the Sun at the very same rate. It's the space-age extension of the leaning tower experiment.

That was Bender's first taste of experimental relativity. His dinner with Weiss in 1974 launched a far more serious involvement. Their conversation about the Huntsville proposal for a massive space-based

interferometer soon expanded in subsequent weeks to include both Faller and Ron Drever. “We had recognized that the ends of the interferometer should be separated as far as possible—that you didn't need a fixed structure,” explains Bender. “We finally got up our nerve to talk about a thousand kilometers or so, without realizing that others had mentioned such a distance with separate spacecraft considerably earlier.” In the early 1970s a number of scientists, including Forward and his colleagues at Hughes, theorists Press and Thorne, as well as Braginsky in the Soviet Union, remarked in print on the possibility of taking interferometers into space as separated spacecraft, not long after Weiss had put his first laser interferometer design down on paper. “Then Ron said, ‘Why stop there?' You just gain by making it longer. We ultimately concluded that a distance of a million kilometers was reasonable to think about it,” recalls Bender. Here was the kernel of an idea that would evolve and gestate for more than 20 years.

Two decades ago the idea of launching a gravity wave detector into space seemed almost fanciful. It was partly hampered by the impression that laser interferometer technology first needed to be developed on the ground. At the time, stable lasers had relatively short lifetimes and low power. With no improvements, the arms of a space interferometer spanning some 1 million kilometers would have had to match within 10 meters. Such fine-tuning would have required frequent adjustments to the spacecrafts' positions, interruptions that would make measurements far more difficult. But once the idea was on the table, interested researchers began thinking of ways around these obstacles. Faller, who gave the first public talk about the concept in 1981, came up with a way to subtract out the laser noise. Bender, using his expertise in celestial mechanics from the lunar ranging experiment, worked out the best orbit—a heliocentric path about the Sun.

Faller and Bender kept the idea alive, aided by modest grants from both NASA and the National Bureau of Standards (now the National Institute of Standards and Technology). This money enabled R. Tucker Stebbins, who had been involved in experimental relativity since his undergraduate days, to join Bender and Faller to put a more sustained effort into the proposal. They even gave their proposed space detector a name: LAGOS, for Laser Gravitational-wave Observatory in Space.

By 1989 LAGOS received high marks from a NASA committee looking into possible space ventures for astronomy after the completion of its Great Observatories program. But when NASA's advanced research funding got cut a few years later, interest turned tepid. “There was even a joking comment about putting us out of our misery,” says Stebbins.

The idea revived, however, when a number of American space interferometer veterans joined forces with a larger group of European gravity wave specialists, including GEO 600's Karsten Danzmann, Jim Hough, and Bernard Schutz. Their formal proposal for a mission, newly dubbed LISA for Laser Interferometer Space Antenna, was presented to the European Space Agency (ESA) in 1993. The project was eventually accepted as an ESA “cornerstone mission,” which will fly once the funds are available. The total cost is expected to be roughly $500 million. The ESA decision, though, was made with the expectation that NASA would eventually come on board as an equal collaborator. NASA is now taking a hard look at the proposal. A LISA project office has been set up at JPL under the guidance of William Folkner, while LISA science teams have been organized on both sides of the Atlantic. LISA is a strong candidate for NASA's 2005-2010 time frame, a period when the ground-based observatories might be detecting their first gravity waves. If approved, it would be part of NASA's mission to explore the structure and evolution of the universe, similar to past agendas that supported such astronomical observatories as the Hubble Space Telescope and the Chandra X-ray Observatory.

LISA was not the sole contender for consideration by NASA and ESA authorities. For a while another idea was circulating within the agencies, a proposal championed by Ron Hellings of NASA's Jet Propulsion Agency. First called LINE, then SAGITTARIUS, and then OMEGA, this plan proposed to put the laser interferometer system in orbit around the Earth rather than the Sun, which makes some aspects of the project easier, such as the launch and telecommunications. In his fifties, Hellings was eager to get something up quickly before he retired. OMEGA called for six spacecraft orbiting the Earth, using existing space platforms rather than special construction. It was to be faster and cheaper, adopting NASA's latest philosophy. The downside

was that a geocentric orbit placed the spacecraft in a more severe space environment. There would be added thermal, geomagnetic, and gravitational forces near the Earth, which would buffet the satellites around. LISA supporters prefer a heliocentric orbit, which allows the spacecraft to maintain a fixed attitude to the Sun, simplifying the forces on the masses. For that and other reasons, many working on OMEGA eventually joined the LISA effort.

Before ESA stepped in, the LISA project had partly been a labor of love, with various participants working on it in their spare time and using their own discretionary funds to finance the initial studies. But once ESA as well as JPL provided seed money, “many nooks and crannies of the design were looked at for the first time,” says Stebbins. The current design calls for three spacecraft to fly in a triangular formation, with the center of the triangle tracing Earth's orbit. The entire system will perpetually follow the Earth like a faithful companion at a distance of some 50 million kilometers (about 30 million miles). Being so far out in space, there will be no seismic disturbances. The test masses, polished platinum-gold cubes 1½ inches wide, will be free-falling along a space-time pathway carved out by the Sun. The orbits were chosen so that each spacecraft will have solar illumination from a constant direction, which will help maintain a stable thermal environment. Once properly inserted into orbit, the three spacecraft should fly in formation with little adjustment, with just a very weak and steady thrust to counter solar radiation pressure, which would otherwise push the spacecraft like wind powering a sailboat.

The three spacecraft will be separated by five million kilometers (3 million miles). Each will be carrying two lasers and two test masses arranged in the form of a Y, so that each spacecraft can be aimed at the other two. This will allow laser light to be continually transmitted and received along each arm of the triangle. Because of the long distances involved, LISA requires a different approach to interferometry. As the laser light travels the long distance from spacecraft to spacecraft, the beam will get wider and wider, eventually spreading out some dozen miles. The original half watt of power will diminish to less than a billionth of a watt at the point of arrival. So the signal cannot simply be reflected back. Rather, the beam must first be amplified—the signal

The three LISA spacecraft in formation. Set up in this traingular pattern, each side 3 million miles long, LISA will follow the Earth in its orbit around the Sun.

boosted—with the onboard laser before being sent back along the arm. If the beam were simply reflected, only a few photons per second would make it back, which would make measurement impossible. Such an amplification scheme is already used in tracking spacecraft but with radio waves instead of laser beams.

Other engineering needs for LISA require sensitivities and specifications not yet fully developed or flight tested. Most important, there

must be a means for maintaining a “drag-free” environment so that all forces upon the test masses (besides solar and planetary gravitational forces) are nearly completely eliminated. Each test mass will be in free fall, isolated as if it were a separate body floating freely in space, so the walls housing each mass must never touch it. Over any one second each chamber must not move in relation to its test mass by more than several nanometers. That' s an extremely small distance, a span just a few atoms thick. The technology to conduct such maneuvers has been used on other satellites but not to the fine performance needed by LISA. The adjustments will have to be made with microthrusters, which might use such means as energized metal ions for the propellant. “Essentially we need the most gentle rocket you can imagine,” says Stebbins. Moreover, they have to worry about such things as the occasional micrometeoroid strike, which would add a sizable push to the spacecraft. If a dust particle, just a few thousandths of a inch wide, hit one of the spacecraft, it would disturb the test mass enough that the microthrusters would have to turn on to maximum thrust temporarily to correct for the shift. The mission is expected to last from 3 to 10 years. If nothing else fails, it will end when the microthrusters release their last little puff of atoms.

LISA would be a complement rather than a competitor to the ground-based interferometers. That's because it would be receiving very low-frequency gravity waves, from a few millionths of a hertz to 1 hertz, well below the band accessible by LIGO and VIRGO. LISA will be detecting the huge long swells in the ocean of space-time, while LIGO and its related kin observe the finer ripples. Each type of wave is generated by either a different astronomical source or a different moment of an event. LIGO and VIRGO, for example, are best tuned to see black hole/black hole binaries when each hole weighs up to a few dozen solar masses. LISA, on the other hand, will be able to spot black hole systems in the range of 100 to 100 million solar masses. Neutron star binaries, on the other hand, will be visible to both but at different times. LISA will see them as two stars orbiting one another, thousands of years before they collide. LIGO will observe them right before the collision, as the stars rapidly spiral inward and the gravity waves sweep to higher and higher frequencies. Thus, instruments are needed both

in space and on the ground to cover the entire spectrum of waves likely to journey through the heavens. A typical wave for LISA, though, will be rather lumbering, taking perhaps 1,000 seconds or more for just one wave to pass by from peak to peak. This is the type of astronomy for those with a patient temperament. Consequently, LISA will ultimately gather fewer data than detectors on the ground. LISA's data rate will be below 1,000 bits a second, versus 6 million bits a second in LIGO. One CD-ROM will be able to store the data from the entire mission.

There is one reason—and one reason only—that supporters have stuck with the idea of a space interferometer through thick and thin financial times. “LISA's greatest strength is its science,” says Stebbins. If the technology works, LISA observers are guaranteed to see something. They might even beat the earth-based interferometers in detecting the first authentic gravity wave. From the ground there are many uncertainties about an interferometer' s ability to see the sources and how many events will ultimately be observable. Ground-based instruments will be working on the margins of detectability. LIGO should see neutron star binaries coalescing, but the event rate is iffy. While supernovas are assured to go off, the strength of the resulting wave is not well known. LISA, on the other hand, will be overwhelmed by sources, inundated by a fairly noisy background. Galactic binaries, such as the myriad number of white dwarf stars orbiting one another throughout our galaxy, will be broadcasting an unremitting cacophony of waves discernible from space. These sources are considered so surefire that “if LISA would not detect the gravitational waves from known binaries with the intensity . . . predicted by General Relativity,” reported a LISA study team, “it will shake the very foundations of gravitational physics.”

The expectation is that, once these binary signals from our galaxy are examined and understood, they can be subtracted from the LISA data. “What you're left with is a record that should have the extragalactic information in it,” says Bender. That would be the signals from massive black holes in faraway galaxies, poised for collision. LISA would be the instrument of choice for studying such supermassive black holes in distant galaxies. These would be gravity wave astron-

omy's most powerful sources. Their examination is one of LISA's prime objectives. One of the best routes for seeing the signals from supermassive black holes, gargantuan objects containing the mass of millions of suns, will be catching two galaxies in the act of merging. During this process, the black holes in each galactic core would eventually coalesce to form an even bigger hole, resulting in a tremendous burst of gravitational radiation. LISA should be capable of seeing the entire last year of the inspiraling holes before their fateful collision. Douglas Richstone, a longtime black hole hunter, calls this “the brightest object in the sky that has not yet been seen.” LISA has the potential to detect 10 of these events each year. It would be sensitive enough to register waves arriving from as far out as 9 billion light-years, across nearly the entire visible universe.

The chances are quite high that LISA will see something, for evidence has been emerging in recent years that all galaxies with a spheroidal bulge, such as elliptical galaxies and most spiraling galaxies, harbor a supermassive black hole in their centers. The larger the bulge, the larger the black hole. Our own Milky Way galaxy harbors a black hole of some 2 million solar masses smack dab in its center. Evidence is mounting that these holes are “quasar fossils,” the engines that once allowed each galaxy when it was young to blaze away with the brilliance of up to a trillion suns. It suggests that the formation of a galaxy and the growth of its central black hole are intimately linked in some as yet undetermined way. Perhaps a modest black hole forms first, serving as the gravitational “seed ” around which a galaxy forms. Over time such a black hole would consume much of the young galaxy's rich supply of stars and gas, bulking itself up. The hole becomes supermassive. Or perhaps the early universe generated a bevy of smaller black holes, each in an individual galactic building block. These separate pieces could then have eventually merged to form a full-blown galaxy, while the black holes coalesced to form a gargantuan black hole in its center.

In either case, gigantic black holes appear to be the natural result of galactic evolution and serve to power the fireworks display that announces each galaxy's birth. This happens because the massive black hole gravitationally attracts any matter lurking near it and never lets it

go. But before this material is permanently captured, it gathers into a swirling accretion disk that surrounds the black hole and radiates intensely. At the same time the hole also spins, like a giant electromagnetic dynamo, producing two jets of subatomic particles that shoot away in opposite directions from each pole at near the speed of light. All of this activity occurs as long as there is enough fuel nearby—stars, dust, and gas—to feed the dark monster in the middle.

Today, quiescent holes can be reignited when galaxies collide. And there are hints in the celestial sky that mergers are in progress. The radio galaxy 3C75, for example, has a set of curving radio jets that resemble a giant water sprinkler at work. These jets appear to be emanating from two nuclei, each possibly an immense black hole. The long spouts get twisted as the two black holes orbit one another. These particular black holes won't be merging for many, many years. But when they at last approach one another, they will emit a distinctive gravity wave signal. The frequency will start low and sweep to higher and higher frequencies as the year progresses and the holes get closer. It will be gravity wave astronomy's ultimate payoff: conclusive proof that black holes are indeed the engines of a galaxy's central activity. LISA could serve as an early warning system. If it can precisely pinpoint the location of an active inspiraling, optical, x-ray, and gamma-ray detectors could be trained on that region to record the final collision.

Almost as interesting will be watching smaller objects—neutron stars, smaller black holes, white dwarfs, and ordinary stars —falling into a supermassive black hole at a galaxy's center. Since stars are so plentiful, this could occur fairly frequently. The final orbits of the doomed stars would be intricate and complex. In some cases, the orbital decay could go on for 70 to 100 years, allowing LISA astronomers to watch them for years. Each object would serve as an exquisite probe for mapping the space-time geometry, the gravitational twists and turns, around a supermassive black hole.

All the while LISA will be listening to that cacophony of gravity wave signals arriving from the host of binary star systems in our own Milky Way galaxy: the waves continually emitted as neutron stars orbit neutron stars, black holes circle black holes, white dwarfs pair up with

other white dwarfs, and all the possible combinations in between. Many of these systems cannot be seen with regular telescopes, so gravity wave detectors will at last offer the means of taking a reliable census of these binaries. Astronomers will hit the jackpot if, in one of those systems, a white dwarf merges with another white dwarf, generating a spectacular supernova. By one estimate, LISA has a 2 percent chance of seeing such an event over its lifetime.

What many scientists look forward to are the waves arriving from events even more bizarre. It is a space-based interferometer that will offer astronomers a particularly good opportunity to look back to the universe's origins, farther than with any other astronomical means. Currently, the cosmic microwave background says something about the universe's condition nearly half a million years after the Big Bang. That's when the primordial fog lifted and the universe became transparent. By then the universe had cooled down enough for neutral atoms to form, at last allowing radiation to travel unimpeded. These light waves, mostly in the optical and infrared regions of the spectrum by that point, were gradually stretched with the universe 's expansion until today they are detected as a vast sea of microwaves. Before that decisive moment in our universe's history, the primordial fireball was a murky soup—a jumble of protons, simple nuclei, electrons, neutrinos, and photons of electromagnetic radiation, all intermixed. Even if astronomers were someday able to peer back to this epoch, they wouldn't see much, for the cosmic plasma was quite opaque, just as the Sun's hot outer layers prevent us from gazing into its nuclear core. This so-called fireball would be an impenetrable barrier to our view.

But a very sensitive space-based interferometer (if not LISA then its successor) could potentially look back as far as the first hundred trillionth of a second after the Big Bang. Unlike electromagnetic waves, gravity waves can cut right through the foggy layers of the primordial fireball. Moreover, there could be other gravity wave events at special moments in the universe's birth, when the universe experienced an abrupt change in its environment—a sudden shift between two different states, like liquid water turning into something far different, ice. As the universe expanded, it too cooled and likely underwent some kind of “phase transitions.” Additional gravity waves might have been emit-

ted in the process. The rapid transition could have given rise to odd defects, areas that retain features of the earlier higher-energy state. That would be the origin of cosmic string. Such strings would be continually oscillating, wiggling, linking up, creating loops. And as they moved, they would generate tremendous wakes, wakes that generate a new supply of gravity waves. Gradually losing energy as they wiggle, they'd finally rocket away in their death throes, disappearing with a final blast of gravitational radiation.

It was surprising, seemingly minor, problems that caused intermittent delays in LIGO's start-up. At Hanford a few weeks had to be spent replacing a glue that failed, the glue that attached the tiny maneuvering magnets to the mirrors. Will LIGO always be at the mercy of such minute details? Fred Raab is not discouraged. “It's like asking the Wright Brothers why they couldn't go up for more than a few minutes at first,” he responds.

While the magnets are being reglued at Hanford, Weiss comes out for a visit to measure any residues remaining after the arms had undergone their “bake-out,” which cleared the steel tubes of their last remaining gases. An electric current had been sent through the beam tubes as if they were long wires. For about a month the arms were heated in this way to 300° F. The total electric bill was over $60,000. Weiss sets up his temporary office in a small portable trailer, parked right outside Port 5, an access door about halfway down the northern arm. He opens a series of valves to link his gas detector to the tube's interior. He patiently sits at a computer in the cramped quarters keenly watching the screen as a graph gradually displays the level of residual gases left in the tube. The first results are encouraging. “Isn't that pretty, ” he says, as the program draws its lines. “I don't see a lot from the tube.” It was too good to be true, though. A half hour later he notices a subtle change in a curve on his computer screen. It's the distinct signature of a leak. Thankfully, a test the next day confirms the leak was occurring at a port rather than a failed weld, a far more difficult repair. A port leak can easily be fixed by tightening a bolt or replacing a gasket.

Weiss gets a bit nostalgic sitting in the trailer. Nearly three decades

have passed since he first sat down and conceived of LIGO's basic design. Now it finally stretches out before him as miles of steel and concrete. “I was here in 1997,” he recalls, “when the first beam tubes were being installed. The meadowlarks and magpies would gather right outside. We'd also see swallows fly straight down the tubes, riding on the thermals. ” After a pause he continues. “A lot of heartache went into this, but it was all worthwhile.”

In his settlement with Caltech over his dismissal from LIGO, Drever chose to develop an independent research program for gravity wave detection. He set up his own 40-meter interferometer for tests. He seems more at ease these days, working once again in his own laboratory. “I like to try out slightly crazy ideas,” he admits. “LIGO is regimented, which is needed to get it built. But my feelings about that are both good and bad. It only allows for ideas that are guaranteed to work, so its sensitivity is marginal. Whether it will see anything is a toss-up, but the potential is there.” His aim is to make the breakthroughs that will guarantee success. For the moment he is working on ideas that will make an interferometer quieter, so that it can be pushed to lower and lower frequencies. “My new lab is for exploring.”

His work space is located in what is known on campus as the Synchrotron Laboratory, because a synchrotron—a type of particle accelerator—was once located in the vast cavernous room. Here Drever has set up his interferometer along two sides of the hall. One arm runs the length of the building. The other arm goes through the width of the building and out the other end, where it continues in a tunnel beneath a road. With this detector he has been trying out a new means of suspending the test masses—magnetic levitation. His hope is that by doing away with the suspension wires he can eliminate a major source of noise. Of course, he has to worry about new sources of noise, such as magnetic field disturbances, but he is hopeful. He and an assistant have a prototype working; a small cube hovers without visible support above an optical table, as if by magic. So delicate is the cube's balance that walking nearby causes a tilt in the floor that makes the mass move forward.

By pushing to lower frequencies, Drever envisions his interferom-

eter being used for more than gravity wave searches. “We might see some interesting geophysics as well,” he says, such as motions within the center of the Earth. Oscillations of the solid core, surrounded by more fluid layers, should generate gravity gradients that theoretically could be picked up at very low frequencies.

“The people who do many wrong experiments are the best,” says Drever. “You try out more things. That's how you make discoveries, but you have to be quick.” There is certainly a quickness to Drever's step, as well as in his speech and in his hands. He's always buzzing around, always thinking, always animated. On this particular afternoon he is most excited by an experiment he had conducted the day before, which caused him to stay in the lab past midnight. He took an old phonograph and used its needle and amplifier to make a crude measurement of the thermal noise in a new material he had just gotten a hold of from another Caltech lab, a material that he says could be a breakthrough for LIGO as a test mass.

The history of the hall where Drever now works goes back much farther than the synchrotron. The immense space, windowless and insulated from the heat of the Sun, was actually constructed in the 1930s to polish the massive 200-inch mirror of the famous Hale telescope, which has been dutifully and majestically surveying the heavens for more than half a century atop Palomar mountain northeast of San Diego. When first conceived by astronomer George Hale, though, the telescope 's success was hardly assured. As at LIGO's conception, much of the technology was not yet developed to guarantee that its mission could be achieved. Many astronomers at the time were highly skeptical that such a large disk of glass—twice the diameter of Mount Wilson's 100-inch telescope that discovered galaxies and the expanding universe—could be poured or properly mounted and maneuvered. The Bureau of Standards had judged that a telescope larger than 100 inches was a technical impossibility. Polishing its thick slab of Pyrex glass has been described as the Apollo project of the Great Depression era. Just several feet from Drever's detector once stood a gigantic turntable upon which the disk—then the world's largest monolithic piece of glass—sat. From 1936 to 1947 (with World War II imposing an interruption), the turntable rotated while a polishing tool pressed

against the disk. Slowly and inexorably the surface was smoothed to exquisite precision while visitors watched from a glass-walled public gallery perched high above the floor. Truckloads of abrasive and jeweler's rouge were used over the years to scrap away tons of glass and sculpt the disk into a paraboloid within 2/1,000,000 of an inch of perfection. Such patience and care won the day. Operating since 1948, the Hale telescope still remains one of the most useful optical telescopes on Earth, even though its once-record size has since been surpassed. As soon as the glass disk was coated with a few grams of aluminum, just 1,000 atoms thick, to turn it into a reflective mirror, the giant concave surface enabled astronomers to peer farther out into the universe than ever before. It helped reveal that quasars were among the first galactic lights to turn on at the dawn of time.

With the Hale telescope successfully put to work, the Caltech optical lab was eventually converted for another use. Instrumentation was erected to gaze inward, down into the inner workings of the atomic nucleus. But now with the synchrotron removed, part of the space returns to its original purpose: perfecting the technology to once again look outward into the cosmos. This time, though, the business at hand is not devoted to a mirror that will gather light waves. Rather, it is to assist physicists in their quest to place their ears on the fabric of space-time and listen to its distinctive sounds.

At first a few notes will register. In time this should lead to a melody that eventually swells into a lush resounding symphony. And when this happens, astronomers will at last be able to discern the hidden rhythms of the universe.