LOS ANGELES -- Frail and dying at the age of 85, John Knudsen Northrop was hardly more than a brown suit full of bones as he shambled into a defense plant conference room on a January day in 1981. The occasion was both sentimental and deeply classified: The Air Force had agreed to make Jack Northrop privy to one of the nation's most secret military ventures, a revolutionary bomber soon to be code-named "Project Senior C.J." The old man signed a secrecy oath and sat down at a table, bracketed by a half dozen executives of the company bearing his name -- Northrop Corp. An Air Force colonel ticked off the proposed bomber's vital statistics: range, speed, payload and, most important, its stealthy ability to elude radar. If successful, the bomber could turn the $250 billion Soviet air defense system into junk, a comforting prospect, given renewed tension between the superpowers. Northrop listened intently, knotting his hands together to control their trembling. Someone hoisted a large wooden box from the floor onto the table. Opening the lid, Northrop saw an 18-inch model shaped like a boomerang, with no tail and only the faintest hint of a fuselage. It bore an uncanny resemblance to the YB-49, a "flying wing" bomber that Jack Northrop had designed and built in the late 1940s only to have the Air Force declare the plane unstable, withdraw its support and order all existing YB-49s chopped into scrap. Heartbroken, Jack Northrop had resigned from his company in 1952, an aerospace genius reduced to self-imposed exile from the industry he had helped create. Northrop's eyes reddened and flooded with tears as he stared at this reincarnation of his flying wing. Choking back a sob, he whispered, "Now I know why God has kept me alive for the last 25 years." A month later, on Feb. 18, 1981, Jack Northrop died. That 18-inch model has become a full-sized bomber with a wingspan of 172 feet. Project Senior C.J. has become the B-2 stealth bomber, by far the most expensive airplane ever readied for production, at more than half a billion dollars per copy. The Air Force wants to build $70 billion worth of them. But the bomber also has become a bat-shaped symbol of the difficult choices facing the nation at a time of colossal budget deficits and a diminishing Soviet threat. To its proponents, the B-2 will help guarantee a reliable nuclear deterrent well into the next century. To its critics, the plane threatens to bankrupt the Pentagon without significantly bolstering the country's security. Congress can't decide what to do about the bomber, underscoring the lack of national consensus on how to arm the American military and how much to spend in the process. For years, Congress monitored the B-2 program with benign neglect, slipping into a tacit alliance with the Pentagon and the defense industry. The B-2 was undertaken with virtually no public debate; the plane's cost, mission and technology were treated as state secrets. Now, the alliance is crumbling, the secrets are emerging from the shadows and, however belatedly, the debate has begun in earnest. If Jack Northrop was the spiritual father of the B-2, William J. Perry was the temporal godfather of stealth technology. A trim, graceful mathematician with penetrating eyes and a puckish sense of humor -- he once walked into a meeting carrying an empty model airplane stand and announced, "Here's the stealth bomber!" -- Perry left his California firm in early 1977 to serve as President Jimmy Carter's defense under secretary for research and engineering. After a month in office, Perry reviewed the four or five most promising technologies being developed by the Defense Advanced Research Projects Agency (DARPA). He immediately latched onto a loosely connected series of experiments known as the "stealth program." If nurtured properly, Perry realized, stealth held enormous consequences for national security. The beauty of stealth lay largely in its simplicity. All moving vehicles disturb the physical world by emitting waves such as infrared radiation from heat, or sound waves, or reflected light, or the reflected radio waves used in radar. These disturbances, called "signatures," can often be detected and tracked by a distant observer. Radar is by far the most prominent signature; as author Bill Sweetman has written, "it is the one that provides an observer with the greatest amount of useful information at the greatest distance." Stealth calls for the suppression or control of all signatures -- particularly radar -- through the shaping and materials used in constructing a vehicle. Since the proliferation of military radar a half century ago, adversaries have desperately sought countermeasures; German U-boat designers in World War II even tried coating submarine conning towers with radar-absorbing materials, or RAM. The United States also had been pursuing stealth -- or "low observable" -- technology; one early enterprise was dubbed Project Harvey, after the invisible rabbit in the Jimmy Stewart comedy. The results, Perry learned in early 1977, had been mixed but promising. By reducing flat surfaces and sharp angles, which are the most egregious reflectors of radar waves, American designers knew it was possible to dramatically reduce the visibility of a plane as measured by its RCS, or radar cross section. The RCS of a B-52 measured from in front of the bomber's nose is about 60 square meters, according to Air Force sources. For the B-1A bomber under construction in 1977, the "nose-on" RCS dropped to about six square meters. For the successor B-1B in the mid-1980s, the RCS declined another order of magnitude to .6; the cross-section of the B-2 is less than .06 square meters, or roughly the radar signature of a small hawk. Among DARPA's stealth projects, Perry learned, was a Lockheed Corp. experiment code-named Have Blue. Using parts from other planes, engineers at the Skunk Works -- the company's secret California plant -- had cobbled together two jets to determine whether they could build a stable, adequately powered fighter while reducing acoustic, infrared and, especially, radar signatures. Both planes eventually crashed in the Nevada desert for mechanical reasons unrelated to stealth. But by the time of the second accident, Have Blue proved that a stealthy airplane could fly and, of equal importance, that computers could accurately predict the signatures of certain aircraft designs. As the Have Blue results came in, Perry moved quickly. He had long worried that complex new weapons inevitably took eight or 10 years to move from brainstorm to battlefield. He thought the process should take only three or four. To streamline the development of stealth technology, Perry created an Executive Committee -- dubbed XCom -- which included key military procurement officials. Appointing himself chairman -- and the only XCom member with a vote -- Perry encouraged the services to submit creative stealth proposals beginning in the fall of 1977. Those judged worthy would be financed generously, with much of the tangle of Pentagon procurement bureaucracy cleared away like so much nettlesome underbrush. Almost immediately, the XCom was swamped with ideas ranging from the clever to the crackpot. "When Lockheed said, 'We think we can make planes that no one can see,' " a retired three-star general recalled recently, "that concept sold like hotcakes in the Pentagon." Every "flying platform" in the arsenal or on the drawing board -- missiles, reconnaissance drones, fighters, bombers -- seemed ripe with stealth possibilities. Some visionaries foresaw even broader applications. If you could lower the RCS of your tank or ship, you had a colossal advantage over any enemy trying to find you. The senior Navy brass was skeptical, however, arguing for "go-slow" caution on grounds that the Soviets were likely to pilfer stealth technology and turn it against the U.S. fleet. That could render obsolete the Aegis radar system, upon which the Navy planned to spend billions of dollars as a mainstay of fleet defense. By 1978, the Navy was working on a stealth missile but the service's initial hesitation was viewed in the Air Force as timid folly -- the "ostrich approach," as one colonel dubbed it.Stealth Goes 'Black' All discussions of the new technology took place within an extremely limited circle. At the urging of National Security Agency Director Bobby Ray Inman, Perry moved all research into the nether world of deepest secrecy. Projects which had been unclassified or simply secret became SAR -- Special Access Required -- a highly controlled, "need-to-know" classification sometimes referred to as the "black world." Among the legacies of Have Blue was a pervasive conviction that the black world was beautiful. Have Blue had been undertaken quickly, competently and at relatively low cost -- the $35 million cost was split between DARPA, the Air Force and Lockheed, according to one official -- without any nattering bureaucrats complaining that the project had run afoul of this picayune regulation or that pointless procedure. To oversee several of the most important stealth projects, the Air Force established a secret "division of low observables" on the fifth floor of the Pentagon's D Ring, in a spare room behind the office that supervised reconnaissance flights. As an indication of the scope and importance envisioned for stealth, Air Force security officers studied historical records to learn how secrecy had been preserved on two previous enterprises: the Manhattan Project, which built the atomic bomb, and the Glomar Explorer, which involved construction of a ship outfitted for ocean mining but secretly operated by the Central Intelligence Agency to recover a sunken Soviet submarine in 1974. The most advanced "L.O." project was a fighter, code-named Senior Trend, which Lockheed began building in its Skunk Works as a full-scale successor to Have Blue. Later named the F-117A, the fighter essentially was an aluminum airplane heavily coated with radar absorbing materials. The jet also was "faceted," or shaped with many small, flat planes intended to deflect radar waves away. Nicknamed the "Wobblin' Goblin" because of early concerns about its stability, the fighter sported a swallowtail, swept-wing look and was intended for deep strikes against enemy air defenses. The plane had its limits, however, including a small payload of bombs, which had to be carried internally to avoid compromising the plane's stealthiness. Essentially a night-flying, fair-weather fighter, the plane also presented maintenance headaches to mechanics who often had to scrape away layers of RAM to service the plane, then carefully glue new RAM back on. A second major stealth project run out of the L.O. Office was a cruise missile, code-named Senior Prom. Also a Lockheed project, Senior Prom found a champion in Anthony R. Battista, an influential member of the House Armed Services Committee's staff who had a technical background and saw to it that initial seed money of $24 million was pushed through Congress in the summer of 1977 to develop this low-flying drone. The third major L.O. project was a bomber. President Carter was not entirely persuaded that the United States needed another one. In June 1977, he had overruled Defense Secretary Harold Brown and killed the B-1A being built by Rockwell International Corp., convinced that it was a waste of money in light of improving Soviet air defenses. But his defense advisers believed that stealth technology could provide a dramatic breakthrough to justify serious consideration of an ATB, or Advanced Technology Bomber, capable of penetrating the Soviet Union for at least a generation. To scientists like Brown -- a one-time protege of physicist Edward Teller -- and Perry, the chance to push forward such an extraordinary new science was immensely exciting, although they viewed the bomber in somewhat different terms. Brown, a Kennedy administration whiz kid who was President Lyndon B. Johnson's Air Force secretary, thought of the plane first and foremost as a nuclear weapon that could be "flushed" -- ordered into an airborne alert pattern -- without necessarily launching an irrevocable attack. For Perry, bombers offered versatility in a nuclear or conventional role because of their ability to carry a large payload long distances. The Air Force could muster countless arguments favoring a new bomber, some persuasive, some specious. One significant factor, most Pentagon civilians agreed, was the service's "silk-scarf" mentality; pilots love to fly, and the dwindling number of large warplanes would shrink even more if a successor to the aging B-52 was not developed. The only substantial opposition came from a faction that preferred to defer the ATB until a resurrected B-1 could be built. "There really wasn't much reason to be against the ATB," one retired general says. "No one ever put in perspective the type of cost we're talking about now. Cost wasn't even discussed much." Lockheed, already at work on the stealth fighter and cruise missile, appeared well-positioned to capture a stealth bomber contract also. But in early 1978, Perry concluded that every flying weapon built for the rest of the century was likely to incorporate stealth technology. He thought it prudent to broaden the American industrial expertise in stealth and encourage some competition among the aerospace companies. No company was more eager to grab a greater share of the stealth business than Northrop. A relatively small firm compared with aerospace giants such as Boeing Co., Northrop was best known for building lightweight fighters. Beginning in the 1960s, the company had done some work in radar cross section predictions, but by 1974 Northrop had all but abandoned this apparently fruitless pursuit since Pentagon research money had dried up. That changed in mid-1975 when a half dozen engineers began working on low observable design with Defense Department encouragement. Northrop submitted a proposed design for DARPA's Have Blue project, but lost the competition to Lockheed. Many Air Force officers admired Northrop's technical virtuosity but felt a visceral dislike for Thomas V. Jones, Northrop's maverick chairman. Jones, an urbane, tireless master salesman who became chief executive in 1960, had been convicted in 1974 of illegal campaign contributions. The Air Force disdain intensified as Jones aggressively -- and unsuccessfully -- sought the service's support for Northrop's F-20 fighter plane, a debacle that eventually cost the company more than $1 billion. At first, the L.O. Office offered Northrop a paltry $50,000 to study stealth. When the company indignantly declined, the Air Force responded with a $13 million contract to develop a bomber concept, an arrangement similar to one made with Lockheed a year earlier. No one thought that Northrop would win the ATB contract -- Perry was willing to offer odds as high as 4 to 1 in favor of the Skunk Works. But the competition, Pentagon officials assured one another, would help keep Lockheed honest.Long Nights in the Desert Yet Northrop possessed something no one had quite anticipated: a small cadre of brilliant engineers with a radically different approach to stealth. Led by electrical engineer John F. Cashen and aeronautical engineers Irving T. Waaland and John Patierno, the company had been hard at work since losing the Have Blue competition. After spinning out theoretical calculations on a computer, the Northrop scientists tested their theories by using small models on an indoor radar range or, preferably, large models at an outdoor range. The most sophisticated testing was done at the government's secret RATSCAT -- Radar Scattering -- range near Alamogordo in the New Mexico desert. On nights when the moon was less than full -- to hide from Soviet spy satellites -- a fiberglass airplane with a 20- or 30-foot wingspan and weighing perhaps 10,000 pounds was hoisted by crane and carefully fastened atop a 60-foot pole as electric lights played eerily across the arid landscape. A radar transmitter half a mile away, using beams carefully scaled to the size of the plane, bombarded the model while engineers took RCS measurements from every conceivable direction to determine which portions of the model -- engine inlets, cockpit canopy and so forth -- gave the most troublesome signature. It was a difficult life. Hugging themselves against the desert cold, the engineers worked quickly to finish before dawn. They worried about the wind toppling the model; they worried about overlooking some critical detail that would fatally flaw a bomber; they worried about spies; they worried about their families. "The security burden was like wearing a lead overcoat all the time. I couldn't tell my wife or friends what I was working on for 15 years," one Northrop engineer recently recalled. "My kids grew up without a father. There's a tremendous personal cost to something like this because we'd be away for days or weeks at a time. The hours were killers. I can't begin to tell you how many desert sunrises I've seen during model testing." In order to have confidence that the RCS readings would correspond to a full-sized bomber, the "black art" of modeling required painstaking exactitude, down to the size of the rivet heads. Northrop found that the best modelers were young men who had built model airplanes as boys or who enjoyed working in automobile body shops. Eventually the company built an average of one model a month, each of which bore a nickname, such as KGB, Manta, Donald and, for the one-eighth scale flying wing, The Hawk, so named for the slight beak shape of the nose which the B-2 displays today. Almost mystical in their thinking, the Northrop engineers considered themselves "phenomenologists" who could "see the waves" in their imaginations as radar beams reflected off a target. By March 1980 -- when John Cashen spent two weeks coaxing a bat-shaped model out of children's clay -- they were committed to a concept very different from Lockheed's faceted approach. The jet would have a central frame of titanium, a strong, lightweight metal that was heat-resistant -- an important consideration since the four engines would be buried inside the aircraft to lessen the chance of an enemy using infrared sensors to detect the bomber. To further lower the infrared signature, the plane would travel at subsonic speeds and engine inlets would be designed with a special tunnel directing cool air into the jet exhaust. In considering various designs, the engineers concluded that Jack Northrop's flying wing was both aerodynamically efficient and, by eliminating the many right angles inherent in a vertical stabilizer, potentially stealthy. Stability concerns, which had plagued the YB-49 in the late 1940s, could be resolved with modern "fly-by-wire" control systems using computers to adjust the pitch and yaw of a jet much faster than a pilot could. The wings would be made of layer upon layer of radar-absorbing graphite tape cemented with epoxy. Portions of the plane, including the leading edge of the wings, would be honeycombed for strength and coated with ferrite-based materials, which could be "tuned" to diffuse various radar wavelengths by conducting the energy in a certain direction. Northrop engineers spoke of "managing" the radar reflection by shaping the plane with compound curves rather than curves with constant radii. Thus, energy which could not be absorbed or attenuated could be guided in directions least likely to be detected. One key attraction of stealth technology was that it reduced reliance on electronic countermeasures (ECM), the center of problems now plaguing the B-1B, to jam or outwit enemy radars. Even so, a contentious debate begun 10 years ago continues to this day over how much ECM the stealth bomber requires. (Northrop has built space for an extra seat in the B-2 cockpit in case a third crewman becomes necessary to run the bomber's electronics). To replicate the "fuzzball" -- the impious term given to natural electronic signals always present in the atmosphere -- considerable attention was paid to technology developed under Project Teal Parrot, which negates some enemy radar frequencies by broadcasting a similar wavelength in a process called "active cancellation." As work on the bomber intensified, Northrop faced a thousand technical challenges. Some were relatively simple, such as adapting smokeless engines, which had become commonplace after the Vietnam War. Eliminating the contrail -- that milky streak of condensed water vapor visible behind high-altitude jets -- was a tougher task and required extensive experimentation with chemicals added to the fuel and exhaust. Even the substantial radar return off a pilot's helmet demanded careful thought, a problem first solved in the F-117A by coating the cockpit canopy with a thin veneer of gold that masked the radar reflection.'We Won It' On Aug. 22, 1980, stealth became public because of political pressures. Against bitter objections from the Air Force L.O. Office, Harold Brown announced that the United States was developing a new technology, which "alters the military balance significantly." The disclosure came in part because only eight members of Congress knew about stealth and the Carter administration had agreed to brief a much wider circle once the technology program eclipsed the $1 billion-a-year mark, which was imminent. Brown assumed that the information would leak and decided to move first. The strategy backfired when Ronald Reagan, Carter's challenger in the November election, accused the president and his defense secretary of crassly exploiting national security secrets. But Democrats, who had been pilloried for opposing the B-1A, could now say they had a strategic weapon they fully supported. Stealth became a Democratic bomber. A few weeks later, in September, the Air Force asked Northrop and Lockheed to submit formal proposals by Dec. 1. The winning contractor was to build 132 planes, with the first squadron operational in 1987. Proposals would be rated on the basis of five general categories -- performance, cost, logistics, management and security. Lockheed's proposal, code-named Senior Peg, dimly resembled a larger, flattened F-117. Northrop's entry, known as Senior Ice, was the now-familiar flying wing. Little love was lost between the two aerospace companies, which had been southern California rivals for decades; one Skunk Works executive sarcastically told the Air Force that Northrop's curvy concept resembled a "Volkswagen bug." Some of the friction reflected a sharp difference in management styles. For example, one of the few members of Congress privy to the bomber competition toured both companies in 1981. At Lockheed's Skunk Works, which resembled a slightly seedy industrial plant, the visitor ate stale doughnuts, drank coffee from a plastic foam cup and was briefed by an engineer who tapped cigarette ashes into his cupped hand as he spoke. Then Northrop. Crisp security guards in starched uniforms greeted the visitor; inside, the plant appeared as tidy and uncluttered as a hospital. Lunch was served on fine china; impeccably dressed executives delivered a well-rehearsed briefing. "There's no doubt in my mind," the departing legislator confided to an Air Force officer, "who could build this airplane." Northrop it was. Lockheed's proposal had been competent, but overall Northrop's entry scored consistently higher. An independent Air Force estimate, according to one source, put the project price tag at roughly $35 billion in 1981 dollars; Northrop's estimate was considerably lower. On Oct. 3, 1981, shortly before Northrop was notified of its victory, President Reagan announced his sweeping plan to modernize the nation's strategic arsenal. Having excoriated the Democrats for canceling the B-1A, Reagan was obliged to resurrect the plane and he did, authorizing a crash program to build 100 of the Rockwell bombers, which would be redubbed the B-1B. In the Pentagon and on Capitol Hill, this "two-bomber" strategy was seen as an opportunity to pit two contractors against one another -- Northrop and Rockwell this time -- while allowing the stealth project to develop at a prudent pace. Air Force Secretary Verne Orr then called Jones on a secure phone line to inform Northrop of its triumph. The bomber, now a major development project, also received a new code-name: Senior C.J., in honor of Connie Jo Kelly, the hard-working secretary in the Low Observable Office (who died of cancer this year). Back in Los Angeles, the exhilarated victors threw a party in the secret warren of offices that the bomber team had rented, improbably, above the Car Barn parking garage on Century Boulevard. Behind the podium they draped a large banner inscribed, "We Won It!" The Pentagon warned Air Force officers to stay away from the fete; careers, egos and huge sums of money had now moved on line, and the relationship between contractor and customer was supposed to remain at arm's length. "It reminded me," one Air Force officer recently recalled, "of a married couple deciding that the time had come to move into separate bedrooms." Word leaked almost immediately that Northrop had won a $200 million initial bomber contract, which eventually could be worth many billions. The company's stock then soared so dramatically that trading was suspended in New York. The Securities and Exchange Commission demanded a prompt explanation; the Air Force sternly warned Northrop to say nothing. Jones flew to Washington, where he labored with several military officers in the L.O. Office to draft a one-paragraph statement that would satisfy the SEC without betraying any more secrets. It was to be the last public comment Northrop would be authorized to make about the project for five years.Staff researchers Melissa Mathis and Lucy Shackelford contributed to this report. NEXT: Building in the black CREATING STEALTH John K. Northrop's "flying wings" were precursors to the B-2 bomber: the piston-driven YB-35, left, and the jet-powered YB-49, below, designed in the late 1940s. After the Air Force withdrew support for the YB-49, Northrop exiled himself from the industry he helped create. THE B-2 DESIGN Stealth is the scientific pursuit of near-invisibility. As applied to the B-2 bomber, stealth involves controlling the "signatures" emitted by an airplane moving through the atmosphere. These include infrared waves from engine heat, sound waves and especially the reflected radio waves used by radar to detect and track an intruding aircraft. Radar signals are "managed" by absorbing, attenuating or rerouting the waves in directions least likely to be reflected to an enemy receiver. For more than a decade, the Northrop Corp. has tried to perfect its stealth concepts through a combination of shape, materials and electronics. A complex sequence of test flights in the next two years will determine whether the B-2 will be as stealthy as Northrop's models and computer calculations indicate. STEALTH'S DISAPPEARING ACT One gauge of an airplane's stealthiness is how much radar energy it reflects. This measurement, known as a "radar cross section," or RCS, is determined largely by the plane's size, shape and materials. Four generations of American bombers have been designed with a dramatically lower RCS, as measured in square meters from the front of the aircraft. ARGUMENTS FOR STEALTH: THE BOMBER'S ABILITY to elude air defenses and inflict massive retaliatory damage helps deter an enemy from attacking the United States. UNLIKE NUCLEAR MISSILES, bombers can be "flushed" -- that is, deployed in an airborne alert pattern without launching an irrevocable attack. Bombers contribute to superpower stability because they are too slow to be considered an effective first-strike weapon. TO COUNTER STEALTH technology, the Soviets would have to consider spending huge sums on air defenses, a diversion of money considered more benign than if it were spent on offensive weaponry. If the Soviets chose not to respond, they would be vulnerable. BY COMPLEMENTING the other two "legs" in the American strategic triad -- intercontinental and submarine-launched ballistic missiles -- the bomber force offers a comforting redundancy that complicates the war-planning if any adversary. BOMBER PILOTS HAVE the ability to determine whether a target has been destroyed by U.S. missiles and, if necessary, can find an alternate target. NEARLY ONE-THIRD of the estimated cost for 132 bombers has been spent. THE B-2 IS THE BEST hope the United States has for eventually being able to find and destroy mobile Soviet missiles and other relocatable targets. UNDER THE REYKJAVIK counting rules drafted by U.S. and Soviet arms negotiators, penetrating bombers like the B-2 are counted as only one warhead, regardless of the actual bomb load. AND THE CASE AGAINST IT: THE ESTIMATED $70 billion cost -- plus many billions more to operate and maintain the bomber -- will bleed money from other critical national needs. BECAUSE OF MANY unanswered questions about the bomber's performance abilities and Northrop's ability to mass produce it, the ultimate B-2 price could be much higher than $70 billion. AN AMERICAN ADVERSARY could discover an inexpensive technological antidote to the B-2, thus sharply devaluing an airplane that costs more than half a billion dollars apiece. IN THE 1990s, quieter Soviet submarines with multiple-warhead missiles may be able to edge close enough to the U.S. coast to destroy most bombers in a sneak attack before they can leave bases. NEW U.S. NUCLEAR weapons -- including an advanced cruise missile, the Trident D5 submarine missile and possibly a small, mobile land-based missile -- will be capable of destroying targets now envisioned for the B-2. THE B-2 CANNOT now find or destroy mobile targets; such capability is unlikely for at least 10 years. INDIVIDUAL STEALTH bombers are so expensive and the technology so sensitive that, contrary to Air Force suggestions, the United States would be very unlikely to use the airplane for military situations other than a World War III. THE RELAXATION of Soviet-American tensions and the growing realization that all-out nuclear war would be catastrophic obviate the need for a new bomber, which adds to the nation's capacity for nuclear overkill.