In 1964 David Wilkinson first went looking for the cosmic microwave background, the compelling evidence that the Big Bang -- the primordial explosion in which the universe is believed to have begun -- really happened. A soft, cold bath of radiation that pervades the universe in all directions, the cosmic microwave background, or CMB, can be detected by a sophisticated receiver tuned to just the right frequency in the microwave band of the electromagnetic spectrum. Wilkinson, at the time an instructor in the physics department, and his colleague Peter Roll set out to build such a receiver, but by the time they had finished it, two radio astronomers at AT&T Bell Laboratories, Arno Penzias and Robert Wilson, had beaten them to the discovery of the CMB.

Penzias and Wilson won the Nobel Prize in physics in 1978, while Wilkinson, who is now the Cyrus Fogg Brackett Professor of Physics, went on to spend his career taking ever more subtle measurements of the background radiation. The pursuit has sent him to many cool, dry places, particularly high ones -- in West Virginia, California, Colorado, Arizona, Hawaii, and Saskatchewan -- where there's less atmosphere between his detection equipment and the depths of the universe.

Through all the hours of all those observations, Wilkinson and his colleagues were learning what the microwave background had to tell them about the early universe from which it comes. Now they know. And what they've learned has made the study of the microwave background arguably the hottest subject in astronomy. The CMB appears to hold the answer to nearly all the favorite questions of astronomers, cosmologists, and physicists. It can reveal the density of the universe, for instance, the speed at which the universe is expanding, and whether it will continue its present expansion forever or eventually slow to a stop and begin to collapse in on itself. The CMB should disclose how much of the universe is made of ordinary matter and how much of a mysterious substance known as dark matter, which has manifested its presence in the gravitational attraction between galaxies but does not appear to emit light. And the CMB is likely to help reveal what that dark matter is. It should help us learn when the stars first turned on, how and why the galaxies formed, and how they clustered together to form the structures we see when we look into space.

All these sublime unknowns seem to have left their mark in the microwave background, although so subtly that the code cannot yet be read. It's as if the makers of the universe wrote down the recipe they used and encoded it in the microwave background. Now all we have to do is measure the CMB with enough accuracy to decipher the recipe, and it will tell us what we want to know (at least until we learn enough to ask new questions).

To that end, Wilkinson and his colleagues at Princeton and NASA's Goddard Space Flight Center in Greenbelt, Maryland, are building a satellite that will measure the CMB to unprecedented accuracy, and NASA has agreed to ante up $70 million to pay for it. The satellite -- known as MAP, for Microwave Anisotropy Probe -- is scheduled for launch in late 2000. According to UCLA astrophysicist Ned Wright, a member of the team working on the project, MAP "will look at the structures of the universe after the Big Bang. And these structures will show us most of the parameters of the universe, what kind of universe we are in. They will tell us the initial conditions out of which galaxies formed."

THE FIRST LIGHT IN THE UNIVERSE

_There comes a time in the study of the universe when sheer unadulterated briliance no longer suffices for progress, and the researchers involved have to get lucky -- preferably, in Wilkinson's words, "lucky as hell." Deciphering the CMB requires that kind of luck. The CMB, which was created by the Big Bang, was the first of two generations of light that exist in this universe. What we see today from the sun and the stars is generation two and was created only after the expanding cloud of the Big Bang had cooled enough for the stars to form.

The first-generation light comes from the Big Bang itself. As the universe expanded and cooled, this original light cooled with it. Today it has cooled to a chilly 2.7 degrees Celsius above absolute zero and can be seen -- or rather, heard with radio antennas -- only when we listen for its soft hiss in the microwave part of the spectrum. That's why it's called the cosmic microwave background.

Perhaps the most telling evidence that these microwaves are indeed the afterglow of the Big Bang is that the radiation is of very nearly equal intensity throughout the universe. It doesn't matter in what direction you turn your antenna, you hear virtually the identical signal at the identical frequency. As Berkeley cosmologist George Smoot puts it, the Big Bang effectively "occurred everywhere simultaneously, hence its afterglow should be uniform across the heavens."

The cosmic microwave background is an extraordinarily soft signal, however, and the universe is a very bright and noisy place because of all that second-generation light and radiation. This is where the luck comes in: the microwave signal is relatively strong and clear at a frequency of 90 billion cycles a second (or 90 gigahertz), which happens to be a frequency at which signal-emitting stars, galaxies, and cosmic dust are most quiescent. "It was just a lucky accident," says Wilkinson. "Nature put its best cosmic information at this frequency and made a wonderful window through our galaxy between the dust and the hot electrons in which to see it."

Wilkinson and his colleagues are not so much interested in simply hearing the hiss as they are in gauging variations in its power from point to point in the sky, a measure known as anisotropy. In effect, the cosmic microwave background has hot spots and cold spots, which differ by all of 1/10,000th of a degree in temperature. Look over here and it might be 2.7281 degrees above absolute zero; look one smidgen of a degree to the left and it might be 2.7282. Such temperature variations reflect density and gravitational fluctuations in the early universe -- hotter spots were less dense than colder spots -- and the current theory has it that those density variations are responsible for the eventual formation of galaxies.

As the universe expanded, explains David Spergel, a professor of astrophysical sciences and a member of the MAP team, the stuff of the universe congregated around those denser regions until eventually there was enough for galaxies to form, then clusters of galaxies and superclusters. The less dense regions, so the theory goes, became the huge voids that can be seen between these superclusters. The discovery of these temperature variations in the microwave background -- the work of a satellite known as the Cosmic Background Explorer, or COBE, in 1991 -- raised two key questions. How did the variations get there? And what do they have to say about the early universe?

As for question number one, cosmologists have two viable hypotheses to explain the origins of the anisotropy. The less favored, known as the defect model, sees the stuff of the early universe, the very fabric of space and time, being laced with cracks and defects as the universe cools. Then, as the universe expands to cosmic size, these defects expand along with it, causing the density and temperature variations and eventually leading to the formation of galaxies.

Cosmologists are not overly enamored with the defect model, says Spergel, because it doesn't fit too well with the data. Most prefer what's known as the inflationary model. In both systems, the universe begins to expand from an inconceivably infinitesimal beginning, a point in space so small it cannot even be adequately described by metaphor. Because the physics of this point is constrained by the laws of quantum mechanics, the energy in it is rife with quantum fluctuations -- ripples not just in the energy density in the point, but in space and time. Within the point, space and time are beset by these quantum fluctuations like an ocean roiling with waves on all scales, from small breakers to deep, unseen tsunamis. When the universe ages by a trillionth of a trillionth of a billionth of a second, it goes through an equally indescribable burp of expansion. It inflates, says Lyman Page, an associate professor of physics, by "some ungodly amount," from something too small to imagine to an emerging fireball at least as big as a grapefruit, if not considerably bigger, and the fluctuations -- these ripples in the energy of the universe -- inflate with it. As the universe continues to grow, so do the fluctuations. "These are seeds," says Page, "and mass can start falling into these fluctuations. And here's the key: it falls in in different ways, depending on what the mass is, and depending on the drag of the mass or how fast the universe is expanding. All those factors enter into it. How it falls into these wells really depends on all these cosmological parameters."

To test the two hypotheses, theoretical physicists take all these crucial parameters, known in the business as the initial conditions -- the density, the speed of expansion, how much matter, what kind of matter, and so on -- and program them into a computer simulation of the expanding universe. "We don't know what the initial conditions are, so we guess those," says Spergel. The computer starts the simulation at a few years after the Big Bang, then evolves and expands the universe for 10 billion to 15 billion years, making sure it dutifully follows the laws of physics. The theorist then examines the simulated universe to see if its galaxies are distributed like the ones observed by astronomers. If so, what patterns are visible in the anisotropy of the cosmic microwave background? How do these density and temperature variations distribute themselves? If the inflationary model is right, the values of all these constants will show up in different patterns of fluctuations in the CMB. If the defect theory is right, then the patterns of those fluctuations will reveal that. All that remains is to measure the CMB with sufficient accuracy and compare that measurement with what the computer simulation predicts. "The idea is to look at the cosmic microwave background and make sure it matches up with a model," says Page. "And to have data good enough to do it."

The catch is getting the data good enough. But since the variations in the CMB are about 1/10,000th of a degree, even the smallest interference -- whether from Earth, the sun, the stars, or the equipment used to detect it -- will overwhelm the variations astronomers are trying to detect. Scientists call this the signal-to-noise problem. The signal is the 1/10,000th-of-a-degree variation in the CMB. The noise is from the instrument itself and all the rest of the universe. For instance, water vapor in the Earth's atmosphere glows with microwaves at the same frequency as the CMB, so you have to listen from someplace very dry. That water vapor will even absorb microwave radiation. If you're not excruciatingly careful, you may think you're measuring variations in the CMB when you're really monitoring the local weather. This is why astrophysicists like to make their measurements from cold sites, because the colder it is, the less water vapor in the atmosphere, and from sites high above the ground (such as high-altitude balloons), because the higher your instrument, the less atmosphere there is to confuse the issue.

Astronomers must do their observing from space in order to map the CMB with sufficient accuracy to measure clearly the variations dating from the first seconds of the universe. The astrophysics community started to suspect this in the early 1990s, which coincided with an idea NASA had for a medium-priced satellite program. It would cost $70 million, and the telescope would be built and launched in perhaps five years instead of the several decades that space projects often take. At that point, the Princeton scientists got together with Chuck Bennett, a NASA astrophysicist, and others at Goddard, the University of Chicago, and UCLA and developed a proposal to build a satellite that would measure variations in the CMB. Two years ago, NASA chose their proposal from some 40 submitted, and the Microwave Anisotropy Probe was in business.

Basically, the satellite consists of a pair of receivers, two metallic horns in the shape of ice cream cones, and amplifiers. The receivers pick up a microwave signal from space. The horns feed the signal to the amplifiers, which boost their volume to a level that's detectable. The signals are then converted to data, which is stored and later transmitted to Earth. The receivers are aimed at points in the sky 141 degrees apart and measure the difference in temperature between them. To work optimally, the receivers have to be cooled to 95 degrees Celsius above absolute zero, which is achieved simply by getting them out into the chill of space. The satellite radiates its heat into space until it is nearly as cold as the surrounding void.

The key to MAP's mission is the spot from which it will be viewing. While its predecessor, COBE, measured the cosmic microwave background from a near-Earth orbit, whizzing overhead at an altitude of 560 miles, MAP will be hanging at a stationary point in space known in the lingo of mathematics and space science as L2, or the second Lagrange point for the Earth-sun system. There are five Lagrange points grouped around Earth. These are the points at which a vehicle will cease to orbit Earth and instead keep perfect time with Earth as Earth orbits the sun. L2, on the far side of Earth a million miles out into space, provides the perfect vantage point for MAP to watch the universe roll by outside its receivers.

MAP will take three months to travel to L2. Once there it will move in tight circles around L2 with its back to Earth. Traveling with Earth in the planet's orbit around the sun, and in its own tiny orbit around L2, MAP will look at each point in the sky for 50 milliseconds, and repeat each measurement a few thousand times over the course of the 27-month mission.

THE RISK FACTOR

It would be nice to say the Princeton- Goddard astrophysicists are sanguine that MAP will work perfectly once it's finished, but that's not quite the case. Their attitude toward the probe is like that of an overly protective mother toward a worrisome child. What could go wrong? "Oh Lord, where do I start?" says Wilkinson. Among other potential problems, MAP is being built on a tight budget, so the probe has very little redundancy: if certain systems fail, there will be no backup systems ready to take over.

Another worrisome possibility is that before MAP even gets to L2, some Earthbound competition will skim the cream off the secrets of the universe. MAP is scheduled for an October 2000 launch. But in 1999, two separate balloon experiments are scheduled to circumnavigate the Antarctic from an altitude of 100,000 feet. They will each have one or two weeks of CMB mapping time, as compared with MAP's 27 months, but that may be enough to learn the basic outlines of what went on in the early universe and make MAP's remarkable cosmic microwave background maps a tad anticlimactic.

This is a distinct possibility, but it stretches the odds for the universe's being a predictable place. Says University of Chicago astrophysicist Steve Meyer, a member of the MAP collaboration, "If the world was as boring as can be and everything is as the theorists predict, then it's very likely many of the things will be known by the time MAP flies. But this would be the first example ever of the world being that boring. If there's something new to learn, these experiments will hint at it. And MAP will show that it's so."

Gary Taubes is a contributing editor at Discover magazine. This article is adapted from a longer version published by Discover last November. © 1997/Gary Taubes. Reprinted with permission.

The Princeton-MAP Connection

Of the 14 members of the science team working on the Microwave Anisotropy Probe (MAP), the satellite that will measure variations in the background radiation permeating the cosmos, 10 have connections to Princeton. On campus are David Wilkinson and David Spergel, professors of physics and astrophysical sciences, respectively, and Mark Halpern, a visiting physicist from the University of British Columbia; Lyman Page, an associate professor of physics and MAP's chief designer; and research physicist Norm Jarosik and postdoctoral fellow Michele Limon, who are building many of MAP's components. Alumni associated with the project are Alan Kogut '67, a physicist at the Goddard Space Flight Center; physicists Stephan Meyer *80 of the University of Chicago and Gregory Tucker *92 of Brown University; and Edward Wollack *94, a postdoctoral fellow at the National Radio Astronomy Observatory, in Virginia.

A member of the team at least in spirit, says Wilkinson, is the late Robert Dicke '39, one of Princeton's greatest physicists, who died last year. It was Dicke who had first proposed that background radiation might exist and who in 1964 set Wilkinson and Peter Roll looking for it. Dicke also invented a device that for years was the standard instrument used for measuring variations in the background radiation.