isa, is a vital port in Roman times, rises 13 feet above sea level on the banks of the Arno river. The city was a sea-going power in the twelfth century, when it became a republic after participating in the First Crusade. At that time its influence extended over the entire coast of Tuscany, sparking economic prosperity and artistic splendor. It was during this period of affluence, in 1174, that Pisa 's campanile—what became the famous leaning tower—went under construction: six tiers of arches on a base, each precariously perched one upon the other and topped with an airy belfry.

The white marble tower resembles an over-tall wedding cake about to topple, as if its builders were on a drunken engineering binge. The lean started as the third level was being built. The ground began to sink due to a water-bearing layer underground. A cultural icon, the structure is continually surrounded by tourists, their cameras clicking away. It is here, the legend goes, that gravity first came to be understood in a scientific manner. Although the story is likely apocryphal, it is said

that Galileo dropped balls of various weights from the top of the campanile to prove his new view of gravity. Until then Aristotle 's word was the standard law of physics when it came to falling objects. The Greek sage had declared that the heavier a mass the quicker it falls to the ground. But from tests he conducted, Galileo concluded that wasn't true. He figured out that the duration of a fall is independent of the mass. Neglecting air resistance, a tiny marble will fall just as fast as a heavy bowling ball. In his Dialogues Concerning Two New Sciences, Galileo had one of his characters, Sagredo, describe the test: “I . . . who have made the test can assure you that a cannon ball weighing one or two hundred pounds, or even more, will not reach the ground by as much as a [hand] span ahead of a musket ball weighing only half a pound.” Galileo begat Newton who begat Einstein. Our modern understanding of gravity began with that one elegant thought.

The leaning tower has been closed off for a while. A series of weights are attached to its base on the north side to counteract the ever-increasing lean. While officials in Pisa work to save the legendary gravitational test site, others nearby are readying a venture to push gravitational research into the future. A French-Italian collaboration is building a LIGO-like detector, known as VIRGO, on the vast alluvial plain just outside Pisa. Although the LIGO detectors are capable of making a gravity wave discovery on their own, since the two separated sites offer the opportunity to reject local disturbances, the biggest scientific payoff will arise when LIGO also operates as an element in an international network of gravity wave detectors. VIRGO will be part of this network, along with other detectors as well. A British-German team has erected GEO 600, a 600-meter-long interferometer, near Hannover, Germany. Japan, meanwhile, has constructed a detector with 300-meter-long arms. Australia also has plans for a large-scale interferometer.

Together, these facilities will eventually form a worldwide system, akin to the global network of radio telescopes that allow radio astronomers to coordinate their observations. Bar detectors are not excluded. Bars and the newer spheres will enhance the network by looking at specific frequencies of the gravity wave signal passing by. When a scientific committee in the United States recommended against the U.S.

building a spherical bar detector, it led to the impression around the world that “bars are out of date,” as one observer put it. But gravity wave researchers are generally agreed that their field needs both types of detectors. “Bars are the only detectors giving us data now,” points out Adalberto Giazotto of the INFN in Italy. “It seems crazy to stop this line of research. We need to push it until there is a detection with both kinds of detectors. Until then it's hard to compare the two techniques. We also have to consider the astrophysics. Maybe the interesting physics will best be seen with spherical bars.” Experts imagine there will eventually be a range of detectors of various designs, each providing a unique contribution toward helping understand the gravity wave signal.

“The search for gravitational waves is a game requiring long, hard effort with a definite risk of total failure,” wrote Kip Thorne in 1980. Two decades later that statement remains as true. Detection depends on them all “listening” and comparing notes. The direction to an abrupt bursting source, such as a binary collision or supernova explosion, cannot be adequately achieved with less than three or four observatories, the way a surveyor needs several points to peg a position. Each site acts like a surveyor's stake in triangulating the location. With LIGO and VIRGO working together, it might be possible to pinpoint a source on the sky to within tens of arcminutes instead of a degree. (A single observatory will still have the opportunity to peg the general direction of a continuous gravity wave source, such as a pulsing neutron star, by noting the Doppler shift in the signal's frequency as the Earth moves in its orbit about the Sun.)

The VIRGO project is a collaborative effort that involves some 100 physicists and engineers from Italy and France. Construction of the interferometer began in May 1996. Its arms are a bit shorter than LIGO's, 3 kilometers (just under 2 miles) instead of 4 kilometers. Set on flat farmland of clay and sand, these arms extend from south to north and from east to west. The surrounding farmers raise sugar beets, corn, and sunflowers for oil. To the north are the mountains from which Michaelangelo obtained the marble for his sculptures. Before laying down the pipes over those many kilometers, VIRGO engineers had to check for lost mines from World War II. One of the

war's most important frontal assaults passed right through this site along the Arno River. The central complex is smaller than LIGO's. Instead of one main building, there are four smaller ones, each constructed out of cement and vinyl siding. From afar it looks like a small industrial park. A lengthy ditch, bordered by earthen dikes, runs by the buildings. This channel serves as an emergency runoff should the nearby Arno overflow. In fact, VIRGO was plagued by frogs during its construction, due to water in its basement.

Giazotto is codirector of the VIRGO project in Italy. His counterpart in France is Alain Brillet. Giazotto's offices are located at an INFN center south of Pisa, in a quiet suburb called San Piero a Grado that is surrounded by a national forest and just a short drive from the VIRGO detector. “Here is where VIRGO was born,” he says with pride as he arrives at the laboratory. The name VIRGO refers to the Virgo cluster of galaxies. The aim is to have an instrument that will be able to detect supernova explosions as far out as that noted collection of galaxies, which is situated some 50 million light-years distant. The hope is that by sweeping over such a vast volume of space the instrument will spot at least a few sources a year. Giazotto is tall and trim, with thinning silver-colored hair and the bearing of an aristocrat, but his office is decidedly utilitarian, with its one desk, one cabinet, and one set of file drawers. Right outside this spartan office is a walkway that looks down on the laboratory where VIRGO detector equipment is being readied. Like many in this field, Giazotto came over from particle physics. An experimentalist, he worked with synchrotrons to study the weak nuclear force and the structure of nuclear particles. But he sees no great difficulty in jumping to the study of gravity. To him it is still a particle physics problem. “To discover the graviton,” he declares is his goal, the theoretical particle that transmits the force of gravity the way a photon transmits the electromagnetic force. “The only way to see them is to build an observatory. And then there's the enormous bonus of better understanding the universe.”

He started thinking about gravitational wave detection in the mid-1970s while he was at CERN working on particle physics experiments. By the next decade he was actively campaigning to get Italy into the business of laser interferometers. But he had a decided point of view.

From the start he wanted to build an instrument that could detect much lower frequencies than the other systems in the works and that meant focusing on the problem of seismic isolation, the largest impediment to detecting low frequencies. He presented the results of his first tests in the mid-1980s at an annual conference on gravity then meeting at the University of Rome. There he met Brillet, a French pioneer in laser interferometry, who was also interested in building a large system. Eventually, they teamed up and arranged for support from the physics communities in both countries. Giazotto firmly believes that detecting lower frequencies is vital for studying certain sources. Take coalescing binary neutron stars, for instance. “From 100 hertz and upward, you have just three seconds of observation time before the stars coalesce,” he notes. “But ifyou start at 10 hertz, you can get a thousand seconds of observing. That's why I want the low frequencies.” His dream is for VIRGO to eventually get down to 4 hertz. In terms of wavelength that would mean gravity waves that span almost 47,000 miles from peak to peak (one-fifth the distance from here to the Moon).

To get to 4 hertz, environmental motions of the suspension system must be reduced by 12 orders of magnitude—down to a trillionth of their original energy. The trick is to stop the outside vibrations from flowing down the wires that hold the masses and thereby jiggle them. To do this, the VIRGO researchers have devised a unique seismic isolation system called the “super attenuator.” It has no rubber cushioning, as in LIGO. Instead, there are six circular rings, stacked one on top of the other to form a structure three stories tall. In certain ways it resembles the multitiered leaning tower, without the lean. Each of these rings is a mechanical filter, consisting of six triangular metal blades under enormous tension, enough to block the noise flowing down the wires to the test mass suspended at the bottom. The scheme appears to work, at least on the prototype erected in the INFN laboratory. “We shook the top with a motor to produce a displacement of about 1 millimeter at 10 hertz,” says Giazotto. “At the bottom we couldn't detect any change, at least to a level of 10^-10.” Ordinary seismic motions, the kind the instrument would face daily, are actually less than that magnitude. Giazotto is very conscious of the high probability of failure, at

least in these first endeavors. “We are really working for those who come after us,” he says.

For VIRGO's future Giazotto favors a controversial proposal: building a second detector fairly close by, within 30 miles or so. He points out that with the LIGO detectors so far apart there is the chance that certain types of waves hitting each one will be “out of phase.” But with two detectors close to one another, the waves would be “in phase,” rising and falling in concert. By adding them together, one would get a decided boost in signal. Of course, there's the added chance that local disturbances would affect both alike, making it difficult to distinguish a true gravity wave. Giazotto believes that a very good isolation system would take care of that concern.

In some ways, VIRGO already has a mate. A second laser interferometer is in place in Europe, the product of one of the world's earliest programs in laser interferometry that has been centered in Germany for a quarter of a century. The German effort in gravity wave detection originated under the guidance of Heinz Billing at the Max-Planck-Institut für Physik und Astrophysik (Max Planck Institute for Physics and Astrophysics) in Munich. Its initial start came almost by accident. At the time the institute included a special physics division devoted to building computers for scientific calculations, a vital need in the days when commercial computers did not yet offer the power to handle complex scientific equations. But as the computer industry caught up to these needs, the institute's computer designers found they were losing their mission. They gained a new vocation when Joe Weber announced that he had detected gravity waves. “The astrophysicists got very excited by this,” recalls Roland Schilling, a former member of the computer-building team. “They said it was so exciting that it would revolutionize all of astrophysics if the claim were true. ” Theorists on staff at the institute wanted to repeat the experiment, but only the computer development group had the necessary laboratory expertise. Almost overnight its members became detector builders for gravity wave astronomy.

The Munich group soon had a room-temperature bar up and running at its facility. The group coordinated its operation with another bar independently built in Frascati, Italy. Within a few years, though,

it became evident that Weber's results could not be confirmed. But a spark had ignited. By then there were two alternatives to improve performance: either construct a supercooled bar or switch to laser interferometry. The Munich researchers, wanting to stay in the business but having neither sufficient expertise nor the necessary infrastructure in cryogenics, opted to go into laser interferometry. They were inspired by the work of both Rai Weiss and Bob Forward. They ordered their first laser in 1974 and were highly optimistic. “I remember very well that we originally had a very short timescale in mind. We thought it would take five or ten years,” says Schilling. It all looked so promising on paper. But just like Drever in Scotland, the Germans quickly discovered the many pitfalls of the new approach. Their first mirrors were rather bad; their poor surface quality scattered the light, which greatly reduced the sensitivity. They spent years learning how to deal with laser beam jitters as well as figuring out the best way to mount the test masses. At first the mirrors were just clamped onto aluminum blocks, but the linkage was a major source of vibration. Their first interferometer was a mere 30 centimeters long. Afterward, they went to 3-meter arms. Despite the technological challenges, they remained encouraged, especially when Joseph Taylor came to a meeting in Munich in 1978 and announced his first gravity wave results from the binary pulsar. Here, at least indirectly, was proof that their elusive goal was not imaginary.

Throughout the 1970s and into the 1980s the German laser interferometer team held all the records for sensitivity for such an instrument. The Munich group got a head start by utilizing Weiss's initial interferometer design and then reducing everything to its bare bones. Along the way they made many valuable contributions to the technique, such as learning to suspend the masses on wire slings to reduce seismic interference and thus let the mirrors themselves serve as the test masses. These innovations are now part of the standard instrumentation in every interferometric gravity wave observatory either planned or built, but at the time the group was immersed in the Model T era of the field, working out the basics now taken for granted.

Schilling has presently invested more than 20 years in gravity wave detection. Since those early days, a time when researchers knew every

worker in the field by name, the community has grown to encompass hundreds of engineers, technicians, astronomers, and physicists, despite the fact that no bona fide signal has yet been captured. “Even if we detect only the sources we expect,” says Schilling, “the scientific gain would be more than worth the effort.” But there is another major reason Schilling is sticking around for the hunt. “If you compare what has been experienced in optical, radio, x-ray, and gamma-ray astronomy, there were always sources that you did not think of before. There were always surprises. Why shouldn't that also hold true for the gravitational wave business? ” he says. It is a sentiment that has become the mantra for the field, its raison d'être.

A turning point for laser interferometry came in 1982. From his work with interferometer control, Schilling came up with the idea for power recycling. At first he thought it wouldn't offer much benefit when incorporated into their design because it required tremendously good mirrors. But Ron Drever was aware that supermirrors were emerging from military technology and independently discovered the same principle: letting the light stay trapped, so that it continues to bounce between the mirrors. This boost makes it appear that the system is using a more powerful laser, which assists in decreasing a major noise in the system. Until this breakthrough, laser interferometry had been the poor cousin to bars in the gravity wave game. Low laser power, scattered-light problems, innumerable vibrations, and poor mirrors made it appear that the technique would remain a dark horse. But by the mid-1980s its status vastly changed. The introduction of power recycling, along with supermirrors and stable lasers, was a critical development. Now laser interferometers are the top contenders, with the bar groups vying to retain a role in a global network of gravity wave detectors. “From the beginning the advantage of the laser interferometers was that they were broadband detectors,” points out Schilling. In other words, they can register a wide band of frequencies, whereas a bar is confined to a narrow frequency range.

In 1983, now as part of the Max-Planck-Institut für Quantenoptik (Max Planck Institute for Quantum Optics) in Garching, a suburb of Munich, the team began to operate a laser interferometer with 30-meter-long arms. After focusing on bettering its sensitivity, the team

switched its emphasis to more technical issues, using the instrument as a test bed for new technologies. “Problems that had to be solved if you wanted to operate a big instrument,” says Schilling. As in the United States, a much larger facility was very much on their minds. At first they thought they would naturally progress up the orders of magnitude. They had already built prototypes with 30-centimeter, 3-meter, and 30-meter arms. It seemed reasonable to try 300 meters. But, politically, they knew they had to jump to 3 kilometers, for that was the size where they'd have a chance to actually detect something, “even though it was against our common sense,” says Schilling. A certain attitude was taking root around the world: let's stop playing with the technology and get some results. At the same time the laser interferometer researchers in Glasgow were pitching the idea for a similar facility to be built in Great Britain. With both countries facing budget crunches in the 1980s, the two groups joined forces in 1989. They named the project GEO, which loosely stood for Gravitational European Observatory. “I said we should call it EGO, for European Gravitational Wave Observatory,” notes Schilling with a smile, a suggestion politely ignored.

In the summer of 1989 Herbert Walther, then managing director of the Max-Planck-Institut für Quantenoptik, met Karsten Danzmann at a laser spectroscopy conference and convinced him to take over the gravitational wave detection effort in Garching. Danzmann had never been involved in gravity wave physics, although he did have a habit of changing fields every few years. Trained in gas discharge physics, he worked for 10 years on heavy ion collisions, laser spectroscopy, and the properties of positronium. He also had experience building ultraviolet lasers. But he readily accepted the invitation to try this new line of research. Gravity wave research takes a person who is “stubborn, will take a risk, and is hopelessly optimistic,” he says in his perfectly accented English, acquired during a stint on the faculty of Stanford University. There is a very subtle boundary, he adds, between genius and being over the edge. “Let's just say that the percentage of people who are on the other side of that edge is a lot higher in this field than in any other,” he says with a laugh. He felt right at home. When Danzmann arrived at Garching, plans for GEO were well under way. But within a

year the project fell apart. The financial difficulties faced by Germany in the aftermath of German reunification made funds for the endeavor disappear. By 1991 support was cut off. Great Britain also froze funding, due to budgetary problems.

Meanwhile, Herbert Welling, a professor of physics at the University of Hannover in Germany, convinced his university to broaden the scope of its research efforts by setting up a chair in gravitational wave physics, just as Caltech had done. Danzmann took the post in 1993, setting up an outpost of the Max Planck Institute for Quantum Optics in Hannover and transferring many on the Garching research team to the northern town. Once in Hannover, Danzmann began talking with James Hough, who had taken over Drever's position as head of the Glasgow gravity wave detection team, to resurrect GEO but on a much smaller scale. “We refused to die,” says Danzmann. They were able to get 10 million marks, less than a tenth of the proposed cost of the original GEO project. “We were sure we could do it for very little money, if we were willing to take risks and do everything in a very unconventional way,” he notes. “We just needed a bit of ingenuity.” They had to be ingenious. The new project was going to be a fifth the size of their original plan.

Because of its 600-meter-long arms (about two-fifths of a mile), the instrument was renamed GEO 600. Its goal is to detect strains of 10^-21, a thousandfold improvement over current prototypes. Danzmann claims they have an advantage in being small. It will enable them to be highly flexible. Equipment can be changed, almost as fast as a new idea is developed, unlike the larger systems where designs must be chosen and locked into years in advance. This allows them to try out interesting new schemes that may well be adopted by the other sites. GEO 600 will act as a testbed for advanced interferometer design. “If you're poor, you have to be smart to survive,” says Danzmann. That means they will have a tiny chance to be the first to detect a signal. Or if signals are first detected with the larger systems, a host of smaller observatories, like the low-cost GEO 600, could be built around the world to enhance the gravity wave network.

GEO 600 is located south of Hannover, about 30 minutes by car. It 's an agricultural test site, land owned by the state government and

operated by Hannover University for agricultural research. The complex was erected right in the middle of working fields of wheat, barley, apples, pears, raspberries, strawberries, and plums. The arms were conveniently built along existing farm roads. Construction was started in the fall of 1995, with a toast of single-malt whiskey, a nod to the Scottish connection. A few drops were sprinkled on the site. “German beer was applied internally later,” says Danzmann. Both Hough and Danzmann, close collaborators for many years, are the principal investigators of GEO 600. Bernard Schutz of the Albert Einstein Institute in Potsdam provides the data analysis expertise.

The size of the instrument was dictated by the width of the land available. It might have been called GEO 573. After 573 meters one arm hits the boundary of the allotted land. To make it an even 600, the collaboration is leasing the last few meters from the farmer next door at a cost of 27 pfennigs per square meter per year. That 's about 270 marks annually. They have no plans to make GEO 600 longer, turning it into a LIGO. Two hundred meters farther down the arm is the Leine River. “It's an experiment,” stresses Danzmann. “It was not meant to last half a century.”

If LIGO were described as an extravagant Broadway show, GEO 600 might be called a high school play. To keep costs down, the German-Scottish collaboration depended heavily on technicians on staff at Hannover as well as student labor. “The entire central building is about as big as LIGO's electronics workshop,” says Danzmann with a rueful smile. “Contractors did the heavy work, like pouring the concrete and putting on the roof, the bare bones building. We did everything else.” Students designed the air-conditioning system for the clean room. It cost 20,000 marks. A commercial system would have cost a million. They also saved money by going with risky designs, such as their vacuum beam tubes, which extend to the north and east. “It's an unproven design that has never been built before. It was rejected by LIGO as being too risky, even though it costs about a tenth of others. We're using a vacuum tube design normally used for air ducts and air-conditioning systems. It's a thin-walled tube, which is stiffened by giving it a corrugation, all the way along. It's like a long bellows. Otherwise, it would collapse. It only has a wall thickness of 0.8 millimeters.

It's a 60-centimeter-wide tube. That way you use very little material, so the material cost alone is low, plus the tube is very lightweight, so handling is very easy. We can use students to carry it around. And the tube supports are a lot less demanding because the weight of the tube is like a wet curtain roughly. Heating of the tube is also easy because the walls are thin. Just a couple hundred amperes of electricity makes it hot,” says Danzmann.

GEO 600 gets its enhanced sensitivity with a promising new technique: signal recycling. Signal recycling was an idea first fully worked out by Glasgow physicist Brian Meers. You might think of it as an interferometer acting like a bar, being “tuned” to a particular frequency. Consider a radio. You tune into a certain station with your dial, which locks your radio onto a carrier wave at a set frequency. The radio then ignores this carrier wave and looks at its “sidebands, ” the frequencies right next door where the music and talk reside. A gravity wave telescope in some ways works similarly. The laser light is a very exact frequency. But if a gravity wave passes by, it moves the mirrors and affects the frequency of the laser light. The signal, in a way, is the “music” placed on either side of the laser light frequency. So the light, as it's being circulated within the interferometer, has these sidebands where the gravity wave signal information resides. In signal recycling it is these sidebands that are stripped off the laser carrier wave and sent back into the interferometer, so that the signal can be built up and amplified.

Despite its shorter arms, GEO 600 has the potential to attain a respectable sensitivity, ideally three or five times less than LIGO over broad bands. And for specific, more narrow frequencies, it may closely approach LIGO and VIRGO. The reason is that GEO 600 incorporates engineering features that the larger systems will be using only later. Along with signal recycling, there is an advanced suspension system. The initial LIGO uses a simple, single-pendulum system with steel wires, which is a proven design. GEO 600, on the other hand, uses a triple-stacked suspension system that employs fused silica wires, a design LIGO intends to adopt in future upgrades. “That's the whole purpose of this instrument,” says Danzmann, “to push the limits of what

you can do experimentally. Of course, if all goes well, it may actually see a signal, although that is not its main purpose.”

Also part of the current worldwide network is Japan's TAMA 300 project. This interferometer is located in Mitaka, 12 miles from Tokyo at the country's National Astronomical Observatory. Unlike the other detectors, the 300-meter-long arms of TAMA are completely underground, housed in long concrete tunnels. There are plans to build a 3-kilometer detector later, which might use mirrors cooled to near absolute zero to reduce thermal noise to a whisper. TAMA 300's shorter arms limit its use as a true observatory, but it is already an important development center and testbed for future interferometer technology.

There was a gap in the layout of the observatories. All the early laser interferometry endeavors were in the northern hemisphere, which greatly restricts the geometry needed to pinpoint the location of a source on the celestial sky. With that in mind, Australian researchers contended that an instrument on their continent would enhance the global network and so actively campaigned to build a detector called AIGO, for Australian International Gravitational Observatory. Almost equidistant from the other instruments, AIGO's location improves both the sensitivity and the resolving power of a gravity wave telescope network. Australia opens up a tremendous volume when it comes to cosmic triangulation. “A southern hemisphere detector is a major component of a global array,” says John Sandeman of the Australian National University. “One interferometer is not really a telescope. An observatory really requires at least a group of four interferometers. ”

Australian researchers have now set up their initial detector on Wallingup Plain, about an hour's drive north of Perth. It's a region well traveled by tourists on their way to Australia's famous Pinnacles. This aborigine land is a sand plain, the perfect medium to absorb seismic vibrations. The interferometer is being built in stages. They have started with a prototype with 80-meter arms to test the technology. Later, as funding allows, the arms will be extended to longer lengths, eventually up to 4 kilometers. This will take its viewing range from the Milky Way out to distant galaxies. Like GEO 600, AIGO scientists will be adventurous in testing advanced materials and designs. “A gravity wave telescope is not like an optical telescope that you can build and

leave for years without change,” says Sandeman. “It's an ongoing process to bring in new technology.” He imagines “digging a hole” into the sensitivity curve of LIGO—in other words, a scheme by which they could go to lower and lower strains, far better than LIGO, but over a limited set of frequencies by using signal recycling. They are also examining the use of artificial sapphire, instead of fused silica, for the mirrors and optics to improve sensitivity.

There is the possibility that when these detectors are turned on worldwide for the first time—the lasers activated and data collected —something will be seen right away. More likely, though, it will be a shakedown cruise, a slow learning experience to understand the instrumentation and its various noises. Most are betting that a signal will not be obtained until the second or third generation of equipment is installed later this decade. But when that signal is snared, it will be a sure-fire candidate for a Nobel Prize, and that lure has the potential to turn the field into a high-stakes competition between the various observatories. Weiss is concerned. “We've seduced people into giving us $300 million for a lark,” he says. “The same thing is happening in Europe and now it's going to happen in Japan and Australia. If the scientists do business as usual, acting like little impresarios, we're in trouble. If it degenerates into that, I'll be very upset.” He wonders whether there will be a carryover from the field of particle physics (where many gravity wave astronomers started out), a mad race to see something before the other guy does.

To prevent such a heated competition, Weiss would like to set up ground rules for confirming a “first detection.” At a meeting of gravity wave researchers in 1998 at the Livingston site, Weiss threw down the gauntlet. “ The heart of the matter,” he says, “will be detector confidence.” For LIGO he would want to have the observation occur in all three LIGO interferometers (the two 4-kilometer instruments and the 2-kilometer system) simultaneously. Moreover, there should be a reasonable delay (the 10 milliseconds of travel time) between Hanford and Livingston, as well as no outstanding environmental interference. All sites should see the same spectra, the same frequency, and the same amplitude. “We're spending a lot of money, so it's crucial to be careful,” he stressed. Weiss imagines that, as the field matures and detec-

tions are more plentiful, the issue of judging a signal will become less volatile. Until that happens, he wants to avoid the conflicts of the past. To the leaders of the world's current gravity wave projects, he offered the following strategy: “A detection of gravitational waves is to be announced only after a statistically meaningful analysis has been performed of the data of ALL instruments that were observing throughout the world. . . . The data and statistical results are brought to a council composed of representatives from each observing group. The initial publication is submitted in two parts. A paper from the group(s) making the observation and their analysis and a second paper from the council discussing the statistical significance in regards to the worldwide effort, in particular, the probability and confidence of detection in some of the instruments as well as the reasons for non-detection in others.”

What if, posed Weiss, one group announces and others disagree? “Then,” he concluded, “we'd have gone full circle to Joe,” referring to the disagreement that continues to this day between Weber and the rest of the gravitational wave community over what his bars were picking up. The community would be elated if gravity wave detectors were set vibrating, while gamma-ray detectors and underground neutrino telescopes registered signals as well from a visible supernova in our galaxy. That would be the Cinderella scenario, but more likely it will be a tough call. As veterans of high-energy physics pointed out to Weiss at the Livingston meeting, it's more realistic to assume there will be leaks to the press about a possible detection, forcing some kind of announcement. In the world of the Internet and fast publication, there's a short time constant on reasoned deliberation. Even the best scientists can have differing opinions on what level of confidence they will settle on to say something is a “discovery” or just “evidence of.” And circumstances can change those assumptions. As one LIGO researcher noted during the debate, “Standards can be lowered as the money runs out.”

LIGO was a project that was close to rejection many times yet survived because of NSF officials who valiantly believed in it and fought to keep it alive. Consequently, there will be great pressure to produce results. “But if LIGO puts out specious results, it'll lose its consensus,”

notes Gary Sanders. To provide extra confidence in a discovery, Weiss would like groups worldwide to refrain from publishing marginal results, especially if one detector sees an alleged signal and others do not. He wants gravity wave astronomers to act locally and think globally. “I look out on all our young people—who now have gray hair. They've been at it for 20 years. It's time we do some science,” he says with great emphasis. “This is more than an experiment. We want to learn something deep about the universe.” There are examples of cooperation in other areas of astronomy. Astronomers have already agreed that any signals that appear to be contacts by extraterrestrials, for instance, must be confirmed by more than one observatory before a public announcement is made. And resonant bar groups already exchange data and share in publications.

But can an observatory be expected to wait for a lengthy time, responded Barish, before it publishes? Is that truly practical? Unless the signal is particularly strong, determining whether a gravity wave has passed by will be a long process. It may take months, even years, to eliminate all the possible sources of interference and reach a confident consensus within just one detector group. It could take a single group a full year to pull a signal out of the noise through elaborate computer processing. Must they then wait another year for another group to painstakingly rake through their data to do the same? It might damage the field more by withholding a claim than by prematurely announcing a discovery. It might be foolish to await some golden event if the project were in danger of being shut down.

Perhaps more troublesome will be the wait itself. Barish expects there will be critics who lose confidence as LIGO and the other detectors work out their bugs over the years. Gravity wave astronomy, at least at its start, will likely require patience. “The first instrument is not the final instrument,” stresses Barish. But despite his technical concerns, he is heartened by the science. “The fact that people can predict gravity wave sources that are within shouting distance makes me feel incredibly confident,” says Barish. “Compared to monopoles, these sources are not just optimistic thinking.”