"I wouldn't want to be quoted on this," President Johnson told a gathering in 1967.
Within NASA, Apollo addressed the concern that Moscow's highly-publicized space spectaculars pointed to communism as the way of the future. The Air Force had a separate space program that dealt with the reality of power. Working closely with the CIA, the Air Force had the task of launching reconnaissance satellites that could determine the Soviet order of battle, counting that nation's bombers and missiles while determining the location of their bases and their operational readiness. In turn, these satellites provided strategic intelligence that shaped America's Cold War policies.
The Air Force in Space
The background to the Air Force program dated to 1953, shortly after the inauguration of President Eisenhower. In August of that year, the Soviets  detonated a nuclear weapon with a yield of 400 kilotons. By studying its fallout, American analysts determined that it was not a true hydrogen bomb. It, however, did represent a large step upward in Soviet nuclear power. In addition to this, the CIA learned that the Soviets were building a turboprop bomber, the Tu-95, with enough range to strike the United States. An intelligence estimate, issued early in 1954, predicted that Moscow would have 500 such bombers in 1957. 2
In March 1954, Eisenhower met with a group of advisors and warned them that he feared a surprise attack, a new Pearl Harbor that would destroy cities rather than battleships. Lee DuBridge, the president of Caltech and chair of this advisory group, he responded by taking steps to set up a high-level commission, the Technological Capabilities Panel. It would recommend new policies that could meet this danger. To chair it, Ike recruited James Killian, the president of MIT.
A subpanel, Project 3, dealt with the technical means for surveillance. The people who learned of it included Clarence "Kelly" Johnson of Lockheed, one of the country's top aircraft designers. He had already prepared a design for a reconnaissance aircraft and, without success, had tried to win support from the Air Force. Johnson now joined with Trevor Gardner, a special assistant to the Air Force secretary, and approached Project 3 for a new try. The subpanel's chairman, Edwin Land, had invented the Polaroid camera and was president of the Polaroid Corporation. He and Killian took the proposal to Ike and convinced him to accept it. The plane that resulted was the U-2. 3
In mid-February 1955, the full Killian Committee issued its report, titled "Meeting the Threat of Surprise Attack." It declared, "We have an offensive advantage but are vulnerable to surprise attack" (emphasis in original). "Because of our vulnerability, the Soviets might be tempted to try an attack." In Edwin Land's section of the report, he wrote,
 At the time, the available "hard facts" were often meager. The 1953 Soviet nuclear test had caught everyone by surprise. Then, on May Day of 1954, at a public air show, the Soviets showed off a new jet bomber, the Bison. Here was another surprise-a Soviet jet bomber. It was all the more worrisome because no one in the U.S. had known of it until the Kremlin displayed it openly. A year later, in preparations for the next such air show, American observers saw a formation of 10 of these aircraft in flight. In mid-July came the real surprise. On Aviation Day, Colonel Charles Taylor, the U.S. air attaché in Moscow, counted no fewer than 28 Bisons as they flew past a review in two groups. This bomber now was obviously in mass production. The CIA promptly estimated that up to 800 Bisons would be in service by 1960.
In fact, Taylor had seen an elaborate hoax. The initial group of 10 Bisons had been real enough. They then had flown out of sight, joined eight more, and this combined formation had made the second flyby. Still, as classified estimates leaked to the press, Senator Stuart Symington, a former Air Force Secretary, demanded hearings and warned the nation of a "bomber gap." The flap forced Ike to build more B-52 bombers than he had planned, and to step up production of fighter aircraft in the bargain. Yet even when analysts discovered the Aviation Day hoax, they took little comfort. If Moscow was trying to fool the CIA, it might mean that the Soviets were putting their real effort into missiles rather than bombers. 5
The U-2 became operational in mid-1956, and proceeded to deliver photos of the highest value. One mission returned with pictures that showed far fewer heavy bombers than expected at Soviet bases. This started a process of downward revision of Moscow's estimated air power. One of Ike's military aides declared that "very quickly we found the Bomber Gap had a tendency to recede. It was something that each year was going to occur. But in fact it did not occur." The U-2 also looked at targets of opportunity, and Richard Bissell, the project manager within the CIA, would recall an example: "He was flying over Turkestan, and off in the distance he saw something that looked quite interesting and that turned out to be the Tyuratam launch site. He came back with the most beautiful photos of this place." It was one of the  principal bases for missile and space launches; yet the CIA had not known of its existence. 6
Nevertheless, the U-2 delivered far less than it had promised. When it entered service, Soviet radar promptly picked it up. Following the second overflight, the Foreign Ministry lodged a protest. The protests escalated, and after only six such missions, all during July 1956, Ike ordered a standdown. Subsequent flights required his personal approval; over the next four years, only about 15 took place. Then, in May 1960, a Soviet antiaircraft missile downed a U-2 near the city of Sverdlovsk. With this, the overflights ceased completely. 7
By then a new concern had arisen: the missile gap. Early in 1960, a debate developed in Washington in response to a new intelligence estimate, which predicted that Moscow would possess up to 450 ICBMs in mid-1963. This would be twice America's anticipated strength in missiles. This was frightening enough; deeper skepticism was raised by the fact that the estimate actually represented a substantial reduction from earlier ones. Senators Symington and Johnson asked whether Ike perhaps was cooking the books, downgrading the perceived threat during an election year. Clearly the nation needed additional strategic reconnaissance, and needed it quickly.
By then, the CIA's Bissell had been working for nearly two years to address this problem. Early in 1958, he had initiated a highly classified program, Corona, that sought to build reconnaissance satellites known as Discoverer. These were to fly to orbit atop Thor-Agena rockets. It took a year and a half, however, to get the system to work successfully. The first attempt, Discoverer 1, did not even reach orbit. Following launch from Vandenberg Air Force Base in February 1959, it wound up near the South Pole. Finally, in August 1960, Discoverer 13 proved the lucky 13 in the series. Though it carried no photo equipment, it successfully demonstrated the release of a capsule from orbit and its recovery in the Pacific. This was the first spacecraft to reenter from orbit and be retrieved following descent by parachute.
With this encouragement, Bissell allowed Discoverer 14 to fly with its camera. Its capsule, too, was recovered successfully, this time in midair, on August 19. The film soon arrived at the CIA's Photographic Interpretation Center, and the photo interpreters gathered in an auditorium. The director,  Arthur Lundahl, spoke to them about "something new and great we've got here." His deputy then presented a map of the Soviet Union. These maps had previously featured a single narrow line to indicate the coverage along the path of a U-2. This one had eight broad swaths running north to south across the USSR and Eastern Europe, covering over one-fifth of their total area. They represented the regions that this single mission had photographed, and people broke out in cheers. Some photos were fogged by electrostatic discharges, but the resolution was 20 to 30 feet, which analysts described as "good to very good." Clearly, this was a turning point. 8
During the election campaign that autumn, Kennedy stressed the issue of the missile gap, warning that the Republicans had done too little to counter its threat. After the election, he appointed a deputy defense secretary, Roswell Gilpatric, who believed strongly that this gap was real. On taking office in January 1961, Gilpatric and his boss, Defense Secretary Robert McNamara, went to the Air Force intelligence office on the fourth floor of the Pentagon and spent several days personally studying Discoverer photographs.
The Air Force held the view that Moscow was building large numbers of well-camouflaged missile sites. Sites for the presumed disguised installations included a Crimean War memorial and a medieval tower. McNamara and Gilpatric, however, preferred the view of Army intelligence: that the Soviet ICBM, designated R-7, was very large and unwieldy and could move only by rail or military road. Discoverer satellites had taken photos along the Soviet Union's railroads and principal highways-and had found no missile launchers. In February, at an off-the-record press conference, a newsman asked about the missile gap. McNamara replied that "there were no signs of a Soviet crash effort to build ICBMs." Reporters raced to their phones, newspapers blossomed with the word that no such gap existed, and Kennedy himself had to step in, declaring that it was too early to draw such conclusions. 9
Then in June and July, Discoverers 25 and 26 flew with nearly complete success. While they were only the third and fourth missions to return photos having intelligence value, together these four flights covered more than half of the regions suitable for ICBM deployment. Within this vast area, photo analysts found no more than two new and previously unsuspected ICBM  bases. Three others were photographed a second time. By comparing them with one another, and with a known testing complex at Tyuratam, the analysts came away with a clear understanding of just what an ICBM base would look like. That made it possible to eliminate a number of "suspect" launch sites and to give a clear and definitive estimate of Moscow's ICBM strength.
This assessment, National Intelligence Estimate 11-8/1-61, titled "Strength and Deployment of Soviet Long Range Ballistic Missile Forces," came out on September 21. It stated:
There indeed was a missile gap-but it favored the United States, and by a large margin. In 1961, the U.S. was already deploying substantial numbers of its first-generation Atlas, Titan, Thor, and Jupiter missiles. In addition, the first Polaris submarines were on station at sea. Beginning in October 1962, the nation would also have the Minuteman ICBM, which would reach the field in even larger numbers. 11
Yet it was hardly a secret in Moscow that the Soviet R-7 was clumsy and unwieldy; that nation's planners had known this from the start. Why, then, had they taken the trouble to develop it? An answer was in hand, courtesy of Oleg Penkovskiy, a colonel in the Chief Intelligence Directorate of the Soviet army. He had recently begun working for MI-6, Britain's intelligence agency, and had gone on to help the CIA as well. In May 1961, he delivered rolls of microfilm that included minutes of Kremlin meetings in which officials decided to use the R-7 for space launches, but not as an ICBM.
 At the outset of the R-7 program, during the mid-1950s, Soviet officials had expected to fire it from secret bases. This missile would take up to 20 hours to fuel and prepare for launch, and during that time, it would be highly vulnerable to attack. However, if the U.S. did not know where these bases were located, the R-7 would remain safe. The advent of American strategic reconnaissance upset this plan, by giving America the intelligence needed to strike bases during pre-launch preparations. The head of the Soviet strategic missile force, Marshal Mitrofan Nedelin, accordingly decided to delay the deployment of a large fleet of ICBMs until he could receive a more advanced version that could be fueled and launched on short notice. 12
The Corona program had sought to use satellites to assess the Soviet threat. It did more; it markedly reduced this threat, at least for a time, by piercing the secrecy that formed a major element of Moscow's strategic calculations. Then, as the 1960s proceeded, the Air Force and CIA, introducing a number of important advances in the satellites, went on to fly them routinely.
Improved resolution was an early goal. The cameras of 1961 achieved a resolution of 25 feet; better versions, in service a year later, reached 10 feet. A further improvement, which first flew in 1967, returned photos with resolution of six feet. To make the best use of these increasingly sharp images, the 1962 and later Discoverer satellites mounted dual cameras that could photograph a site from different directions. This permitted stereophotography, whereby analysts could study images that appeared three-dimensional. Later versions of this spacecraft also carried more film and stayed up longer. The first Discoverers had mission times of a single day; subsequent models stretched this to three weeks and longer.
The six-foot resolution represented a limit for Discoverer. The Thor-Agena booster, used by the Discovere, had only a modest payload capacity. Beginning in 1963, however, the Air Force employed the Atlas-Agena and then the Titan III, which could launch larger spacecraft with telescopes of greater acuity. These rockets supported a separate program, Gambit, that achieved a resolution of 18 inches. For the closest looks, the Air Force used the closest orbits, with Gambit spacecraft dropping down to perigees as low as 76 miles. 13
 What did Corona and Gambit show? They photographed all Soviet ballistic-missile launch complexes, following existing as well as new missiles through development and deployment. In particular, they found and repeatedly observed a major center at Plesetsk, near the northern city of Arkhangelsk. Plesetsk specialized in launching reconnaissance satellites and other military spacecraft. At its height, it accounted for more than half of all space launches in the entire world, with Tyuratam a distant second, and Cape Canaveral and Vandenberg Air Force Base far behind.
Corona also was first to see Severodvinsk, the main construction site for ballistic-missile submarines. This made it possible to monitor the launching of new classes of subs, and to follow them through to operational deployment. The CIA also observed the rapid growth of the Soviet surface navy. Coverage of aircraft plants and air bases kept analysts up-to-date on bombers and fighters, while other coverage allowed Army experts to learn the nature of the tank forces that NATO would face if the Soviets were to invade Europe.
Corona photography uncovered the construction of antiballistic-missile sites near Moscow and Leningrad, along with the radar installations that supported them. Other photos located antiaircraft batteries and made it possible for the Strategic Air Command to find routes for its bombers that could avoid these missiles. Specialized satellites, conducting geodetic mapping, became the main source of data for the military charts of the Defense Mapping Agency.
As recently as the mid-1950s, the Soviets had been able to fool the Americans concerning their air strength, and to touch off a major Washington flap over a supposed "bomber gap," merely by flying the same aircraft around twice at an air show. By contrast, a 1968 intelligence report contained the unequivocal statement: "No new ICBM complexes have been established in the USSR during the past year." As early as June 1964, Corona had photographed all 25 of the complexes then in existence. If there had been any new ones, the CIA would have seen them. 14
The Air Force and NASA
In 1494, the Treaty of Tordesillas divided up the New World by drawing a line down the Atlantic, with Spain claiming lands to the west of this line and  Portugal claiming lands to the east. 15 The activities of NASA and the Air Force lent themselves to similar demarcation. With NASA emphasizing Apollo while the Air Force dealt largely with satellite reconnaissance in low orbit, there was little overlap between their concerns. However, these two agencies did not run independent programs; there was a great deal of cooperation.
This cooperation was particularly strong in the realm of launch vehicles. In launching automated spacecraft, the most important such vehicles were derived from the Thor, Atlas, and Titan ballistic missiles; both NASA and the Air Force used these rockets repeatedly, and procured them from the same contractors. They also shared in ongoing developments that increased their payload capacities.
As early as February 1961, an agreement between NASA's James Webb and the Pentagon's Roswell Gilpatric stipulated that neither agency would initiate the development of a new launch vehicle without first seeking the consent of the other. Then in 1962, a joint NASA-DoD Large Launch Vehicle Planning Group issued a report that contained a recommendation: "The 120-inch diameter solid motor and the Titan III launch vehicle should be developed by the Department of Defense to meet DOD and NASA needs, as appropriate in the payload range of 5000 to 30,000 pounds, low earth orbit equivalent." 16
The Titan III brought the prospect of wasteful duplication, for it competed directly with NASA's Saturn I-B. This Saturn carried over 36,000 pounds to low orbit. The Titan III-C, the first operational version, had a rated payload of 23,000 pounds; its immediate successor, the Titan III-D, raised this to 30,000. In addition to this, the projected Titan III-M promised to carry as much as 38,000. Nevertheless, as early as 1967, the President's Science Advisory Committee noted that "the launch costs of the [Saturn I-B] are about double those of the Titan III-M." 17
Because NASA was accustomed to receiving launch vehicles that the Air Force had developed, it yielded gracefully when the Saturn I-B came under pressure. NASA had conducted the initial flight test of a Saturn-class first stage as early as October 1961, at a time when the Titan III was still at the level of preliminary study. In view of this early start, and because the Saturn....
....I-B was essential for Apollo, NASA went on to build 14 of them, though George Mueller hoped for more as he pursued Apollo Applications. When budget cuts hit home, however, NASA abandoned the Saturn I-B and turned to the Titan III-E Centaur. It had the energy to launch large payloads on missions to Mars and the outer planets, and did so repeatedly. 18
In addition to launch vehicles, NASA turned to the Air Force for facilities used for launch and tracking. When NASA's rockets flew from Cape  Canaveral, they proceeded down the Eastern Test Range-which the Air Force operated. That service provided tracking stations, and when NASA built stations of its own on the islands of Antigua and Ascension, they were co-located near those of the Defense Department.
The Air Force also built up an extensive array of launch facilities at Cape Canaveral. When NASA took over nearly exclusive use of some of them, the Air Force transferred them to NASA outright. These included Launch Complex 12 for Atlas-Agena, LC 36 for Atlas-Centaur, and LC 19 for the Titan II. Other launch pads served both agencies: LC 17 for Delta, LC 41 for Titan III. In addition to this, NASA launched early versions of Saturn, including the Saturn I-B, from LC 34 and 37, which had been built on land owned by the Air Force.
The two agencies also cooperated closely in research. The Air Force had a valuable set of wind tunnels and engine-test facilities at its Arnold Engineering Development Center in Tennessee. This service, however, did not attempt to duplicate the far more extensive facilities of NASA-Ames, Langley, and Lewis. In addition to a broad array of supersonic wind tunnels, NASA offered such unique installations as a wind tunnel at Ames Research Center with a 40 by 80-foot cross section, big enough to hold and test full-size fighter aircraft. At NASA-Langley, a 60-foot vacuum sphere could accommodate large spacecraft and rocket stages. 19
In addition to sharing facilities, NASA and the Air Force also pursued joint ventures in research. The X-15 was one; another, the XB-70, involved large aircraft that could fly at Mach 3. The agencies also collaborated in building immense solid-propellant rockets. At Edwards Air Force Base, NASA built test stands for rocket engines used in Apollo. These complemented earlier Air Force test facilities.
Institutional arrangements also bound them closely. Between 1958 and 1964, NASA and the Defense Department executed some 88 major agreements. A joint Aeronautics and Astronautics Coordinating Board (AACB) dealt with such areas as aeronautical research, launch vehicles, spacecraft, and piloted space flight. NASA's Deputy Administrator and the DoD's Director of Defense Research and Engineering co-chaired this board; as early as 1966, an AACB subpanel carried out an important review of concepts for reusable launch vehicles.
Within the Defense Department, the Air Force Systems Command (AFSC) held overall responsibility for that service's space and missile programs. In downtown Washington, an AFSC liaison office shared a building with NASA's Office of Manned Space Flight (OMSF). After 1962, NASA had its own Office of Defense Affairs that performed a similar function. 20
 Yet, this interagency cooperation would only go so far. In August 1963, Webb and McNamara signed an agreement that sought "to ensure that in the national interest complete coordination is achieved" in pursuing a joint space-station project. Only a month later, McNamara sent Webb a follow-up letter that expressed his reservations. Then in December, McNamara made it clear that at least for the short term, the Air Force would want a piloted orbital facility of its own. 21
When he canceled Dyna-Soar, on December 10, he handed the Air Force a consolation prize by inviting that service to conduct studies of a new project, the Manned Orbiting Laboratory (MOL). MOL took shape as a cylinder, 10 feet across by 41 feet long, with a Gemini spacecraft at one end; this ensemble was to ride to orbit atop a Titan III. McNamara could not grant formal approval for MOL; that had to come from the White House, and this raised anew the question of what the Air Force might do with such a facility. While that service had failed to provide good justification for Dyna-Soar, this time it came up with a fine reason for MOL: strategic reconnaissance.
The eventual plan called for MOL to carry a telescope with aperture of six feet, offering resolution of nine inches. Astronauts would avoid photographing cloud-covered regions, but would scan the ground with binoculars, looking for items of interest. The Air Force won support from such key figures as Kermit Gordon, director of the Budget Bureau, and Donald Hornig, the White House science advisor. In August 1965, President Johnson gave MOL his endorsement, which meant it could go forward to contract award and development.
By then George Mueller was nurturing hopes for Apollo Applications, which raised anew the prospect of duplication. NASA officials, unwilling to affront the Air Force, supported MOL and took the view that it was not the national space station contemplated in the 1963 Webb-McNamara agreement. Nevertheless, members of Congress as well as Budget Bureau officials soon were asking whether NASA could adopt a version of MOL for its own use. 22
In January 1966, Senator Clinton Anderson, chairman of the Senate space committee, sent a letter to Webb that recommended use of MOL. Within the House, the Military Operations Subcommittee criticized Mueller's plan for "unwarranted duplication" and called for MOL to proceed as a joint NASA- Air Force program-with the Air Force in charge. Budget Bureau officials also supported a common program. In February 1967, the President's Science Advisory Committee added its own views, calling for "maximum utilization" of MOL and calling on NASA to carefully consider its use "before substantial funds are committed" to Apollo Applications.
NASA responded by having Douglas Aircraft, the MOL prime contractor, evaluate the suitability of MOL for NASA's objectives. The agency also conducted in-house studies. These began by acknowledging that the Saturn I-B was far more costly than the Titan III, and considered whether it might be advantageous to have the latter launch Apollo spacecraft. The OMSF concluded that while this was possible, it would cost $250 million to develop such a Titan-Apollo, which would then require 17 launches before the savings surpassed the initial cost of conversion.
The OMSF also concluded that MOL was too small for NASA's needs. It was no larger than a house trailer, whereas Mueller had described his proposed wet workshop as being the size of "a small ranch house." While the Air Force had a proposal in hand for a larger MOL, this would cost an additional $480 million and would take four years to develop. In comparison, even the Saturn I-B would cost less to use. These arguments mollified the critics, and Apollo Applications went forward, though with a reduced budget. 23
MOL also went forward, with strong Pentagon support. The Air Force, however, never having carried through the development of a piloted spacecraft, failed to control its cost. Between 1965 and 1969, the projected cost of MOL ballooned from $1.5 billion to $3 billion. During those same years, the escalating Vietnam War placed military programs under severe strain.
The future of MOL came up for discussion at a White House meeting between Nixon, national security advisor Henry Kissinger, and Budget Bureau director Robert Mayo. Though the program carried the strong endorsement of Defense Secretary Melvin Laird and the Joint Chiefs of Staff, it proved to lack support from a key official: Richard Helms, Director of Central Intelligence and head of the CIA. In the words of the analyst Jeffrey Richelson, Helms's advisers "feared that an accident that cost the life of a single astronaut might ground the program for an extended period of time and cripple the reconnaissance program."
 Mayo suggested canceling MOL; Nixon and Kissinger agreed. Only then did Mayo discuss the matter with Laird, who had not even believed that MOL was in trouble. Though Laird appealed directly to Nixon, emphasizing that the Joint Chiefs firmly supported this program, Nixon turned him down. The public announcement of the end of MOL came on June 10, 1969, with its first piloted flight still three years in the future. 24
During the 1960s, the Air Force pursued two major and separate efforts-Dyna-Soar and MOL-that sought to place military astronauts in orbit. This service ended the decade with both projects canceled and with nothing to show for its efforts. Clearly, if it was to send such astronauts aloft, it would not do so on its own, but would have to work in cooperation with NASA.
In addition to this, the experiences of the Titan III, the Saturn I-B, MOL, and Apollo Applications, had shown clearly that these agencies could easily introduce wasteful duplication by pursuing their own programs. This made it plausible that a cooperative NASA-Air Force program, focusing on piloted flight to orbit, would take shape as a national program, a unified effort shaped to serve the needs of both agencies.
Clearly, Air Force involvement would emphasize strategic reconnaissance, which represented the main rationale for that service's activities in space. The experience of MOL, however, showed that it would not do simply to propose that astronauts could operate telescopes and cameras from orbit. Instead, the Air Force would have to use piloted flight to support its work with automated reconnaissance satellites, such as those of Corona and Gambit.
As early as 1963, NASA and the Air Force had executed the Webb-McNamara agreement, which contemplated a joint space station. With MOL now canceled, it was difficult to see how the Pentagon could justify major participation in the space station that NASA's Tom Paine wanted so badly. The Space Shuttle was another matter. By launching, retrieving, and servicing spacecraft, it might significantly enhance the ability of the Air Force to conduct strategic reconnaissance. In turn, by serving Air Force needs, the Shuttle might indeed take shape as a truly national system, carrying military as well as civilian payloads. Beginning in 1969, the evolution of the Space Shuttle concept took a sharp turn in this direction.
 A New Shuttle Configuration
When a new round of shuttle design studies got under way, early in 1969, the field had seen no truly new concept since Max Hunter's partially-reusable Star Clipper of several years earlier. While work went forward at the contractors, Max Faget, at the Manned Spacecraft Center, carried through a parallel effort of his own that indeed came up with a new approach. His configuration not only went on to dominate the alternatives; it changed the terms of the ongoing discussions. These discussions had emphasized such issues as full versus partial reusability, with neither approach finding expression in a generally-accepted design concept. Faget now introduced a specific concept: a two-stage fully-reusable shuttle. As it gained acceptance, it spurred debate over its specific features, notably size, payload capacity, and choice of wing design. By focusing the debate, Faget's work thus narrowed the topics that subsequent studies would address, and enabled these studies to achieve greater depth.
Faget was an aerodynamicist who had built his career at NACA's Langley Aeronautical Laboratory. He was a member of the Pilotless Aircraft Research Division, an early nucleus of activity in high speed flight. In 1954, he took part in an initial feasibility study that led to the X-15. He then found a point of departure for his subsequent career in the findings of his fellow NACA aerodynamicists, H. Julian Allen and Alfred Eggers. They had shown that for a reentering nose cone, a blunt shape would provide the best protection against the heat of reentry.
Working with a longtime associate, Caldwell Johnson, Faget proceeded to devise a suitable blunt shape for Project Mercury, which put America's first astronauts in orbit. His Mercury capsule took shape as a cone, with its broad end forward and covered with a thick layer of material to provide thermal protection. (A cutaway view of this concept, elegantly rendered, hangs in Faget's offices to this day.) He came to Houston as a founding member of the Manned Spacecraft Center (MSC), where he became Director of Research and Engineering. He also adapted his basic shape to provide capsules for Gemini and Apollo. 25
"My history has always been to take the most conservative approach," he declares. In this frame of mind, he disliked much of the work done to date on Space Shuttle concepts. Lifting-body configurations were popular; Lockheed's Max Hunter had used them in his Star Clipper. Faget acknowledged their merits: "You avoid wing-body interference," which brings problems of aerodynamics. "You have a simple structure. And you avoid the weight of wings." He saw difficulties, however, that were so great as to rule out lifting bodies for a practical shuttle design.
They had low lift and high drag, which meant a dangerously high landing speed. As he put it, "I don't think it's charming to come in at 250 knots." Engineers at McDonnell Douglas, studying their Tip Tank lifting body, had tried to improve the landing characteristics by adding small wings that would extend from the body during the final approach. This appeared as very makeshift.
Because they required a fuselage that would do the work of a wing, lifting bodies also promised serious difficulties in development. It would not be possible....
....to solve aerodynamic problems in straightforward ways; the attempted solutions would ramify throughout the entire design. In his words, "They're very difficult to develop, because when you try to solve one more problem, you're creating another problem somewhere else." His colleague Milton Silveira, who went on to head the MSC Shuttle Engineering Office, held a similar view:
Instead, Faget proposed to build each of his shuttle's two stages as a winged airplane, with thermal protection on the underside. Before it could fly as an airplane, such a shuttle would first have to reenter, which meant it  would need the high drag of a blunt body. "With extremely high drag," he notes, "you throw a big shock wave in front of you, and all the energy goes into that shock." Even with thermal protection, he did not want to fly his shuttle during reentry, in the manner of an airplane: "It's a hell of a lot easier to do a no-lift entry than a lifting entry, from the standpoint of heat protection." With airplane-style reentry, "you are stuck in the atmosphere, going fast for a long time." Rather than lose energy to a shock wave, the airplane would experience drag through friction with the atmosphere which would transfer heat to its surface.
Faget expected to turn his airplane into a blunt body by the simple method of having it reenter at a very high angle of attack, with its broad lower surface facing the direction of flight. In effect, he would take an Apollo capsule, with its large circular heat shield, and trim it to the shape of an airplane with wings. This concept drew on the experience of the X-15 that looked like a fighter plane but reduced its reentry heating by coming in nose-high. It also revived a design approach introduced a decade earlier by NASA's Charles Mathews. He had also proposed to build a winged spacecraft as a glider that would reenter with its bottom side facing forward.
Faget wrote that "the vehicle would remain in this flight attitude throughout the entire descent to approximately 40,000 feet, where the velocity will have dropped to less than 300 feet per second. At this point, the nose gets pushed down, and the vehicle dives until it reaches adequate velocity for level flight." This dive would cost some 15,000 feet of altitude. The craft then would approach a runway and land at a moderate 130 knots, half the landing speed of a lifting body.
Faget wrote that because its only real flying would take place during this landing approach, a wing design "can be selected solely on the basis of optimization for subsonic cruise and landing." The wing best suited to this limited purpose would be straight and unswept, like the wings of fighter planes in World War II. A tail would provide directional stability, again as with a conventional airplane. By moving control surfaces on the horizontal stabilizer, a pilot then could raise the nose slightly just before touching down on a runway, in a maneuver called a flare which adds lift and makes the touchdown gentle. 27
 Faget's concept had the beauty of simplicity, and, inevitably, knowledgeable specialists would criticize it as being too simple. The Air Force Flight Dynamics Laboratory (FDL), at Wright-Patterson Air Force Base, quickly emerged as a center of such criticism. The FDL had sponsored space shuttle studies in parallel with those of NASA, and had investigated such concepts as Lockheed's Star Clipper. One of its managers, Charles Cosenza, had been a leader in the development of ASSET. Another FDL scientist, Alfred Draper, was a leader in the field of space systems. Beginning in early 1969, he took the initiative in questioning Faget's approach. 28
Draper did not accept the idea of building a shuttle as an airplane that would come in nose-high, then dive through 15,000 feet to pick up flying speed. With its nose so high, the plane would be fully stalled, and the Air Force disliked both stalls and dives, regarding them as preludes to an out-of-control crash. Draper preferred to have the Shuttle enter its glide while still supersonic, thus maintaining much better control while continuing to avoid aerodynamic heating.
If the Shuttle was to glide across a broad Mach range, from supersonic to subsonic, then it would encounter an important aerodynamic problem: a shift in the wing's center of lift. Although a wing generates lift across its entire lower surface, one may regard this lift as concentrated at a point, the center of lift. At supersonic speeds, this center is located midway down the wing's chord (the distance from leading to trailing edge). At subsonic speeds, this center shifts and moves forward, much closer to the leading edge. Keeping an airplane in proper balance requires the application of an aerodynamic force that can compensate for this shift.
The Air Force had extensive experience with supersonic fighters and bombers that had successfully addressed this problem, maintaining good control and handling characteristics from Mach 3 to touchdown. Particularly for large aircraft-the B-58 and XB-70 bombers, and the SR-71-the preferred solution was a delta wing, triangular in shape. Typically, delta wings ran along much of the length of the fuselage, extending nearly to the tail. Such aircraft dispensed with horizontal stabilizers and relied instead on elevons, control surfaces resembling ailerons set at the wing's trailing edge. Small deflections....
 .....of these elevons then compensated for the shift in the center of lift, maintaining proper trim and balance without imposing excessive drag. 29
Draper proposed that both stages of Faget's shuttle should feature delta wings, rather than straight ones. Faget would have none of this. Though he acknowledged the center-of-lift problem, he expected to avoid it: "The straight wing never flew at those speeds; it fell at those speeds." A delta wing with elevons promised problems at landing, when executing the flare prior to touchdown. That flare was to add lift, but raising the elevons would increase the drag-with the added lift coming only after the nose had time to come up. This momentary rise in drag would make the landing tricky and possibly dangerous.
To achieve a suitably slow landing speed, Faget argued that the delta wing would need a large wingspan. A straight wing, having narrow chord, would be light and would offer relatively little area demanding thermal protection. A delta of the same span, necessary for a moderate landing speed, would be physically much larger than the straight wing. It would add considerable weight, and would greatly increase the area that would receive thermal protection.
Draper responded with his own viewpoint. For a straight wing to deal with the shift in center of lift, a good engineering solution would call for installation of canards, small wings mounted well forward on the fuselage that would deflect to give the desired control. Canards produce lift, and would tend to push the main wings farther to the back. These wings would be well aft from the beginning, for they would support an airplane that was empty of fuel but that had heavy rocket engines at the tail, placing the airplane's center of gravity far to the rear. The wings' center of lift was to coincide closely with this center of gravity. Draper wrote that the addition of canards "will move the wings aft and tend to close the gap between the tail and the wing." The wing shape that fills this gap is the delta; Draper added that "the swept delta would most likely evolve."
The delta also had other advantages as well. Being thick where it joins the fuselage, it would readily offer room for landing gear. Its sharply-swept leading edge meant that a delta would produce less drag than a straight wing near Mach 1. In addition to this, when decelerating through the sound barrier, a delta would shift its center of lift more slowly. The combination of a sudden drag rise near Mach 1, combined with a rapid center-of-lift shift, would  produce a sudden and potentially disconcerting change in the stability characteristics of a straight-winged shuttle when slowing through the speed of sound. This change in stability would be much less pronounced with a delta, and would give a pilot more time to react. 30
The merits of deltas might have remained a matter for specialists for another important feature of the delta: Compared to the straight wing, it produced considerably more lift at hypersonic speeds. Using this lift, a reentering shuttle could achieve a substantial amount of crossrange, flying large distances to the left or right of an initial direction of flight. The Air Force wanted plenty of crossrange, and the reasons involved its activity in strategic reconnaissance. In particular, these reasons drew on recent experience involving the Six-Day War in the Middle East in 1967, and the Soviet invasion of Czechoslovakia in 1968.
The Six-Day War broke out suddenly, pitting Israel against a coalition led by Egypt whose tanks and aircraft came largely from the Soviet Union. Though America's intelligence community sought to follow the fighting closely, its means proved to be limited. A ship near the Israeli coast, the USS Liberty, monitored the communications of the belligerents-until the Israelis bombed it. Though spy planes, such as the U-2 and SR-71, could look down through clear desert skies, experience had shown that the U-2 was vulnerable to antiaircraft missiles. Satellite reconnaissance relied on Gambit and Corona, which had been designed to follow the slow development and deployment of missiles and other strategic weapons. They were not well-suited to the swift battle maneuvers of the 1967 war, and Defense Secretary McNamara does not recall that these spacecraft played any role in U.S. intelligence-gathering during those six days. By the time Corona photography became available, the war was over already.
Then, in August 1968, the Soviets stormed into Prague. A Gambit spacecraft, launched on August 6, performed poorly and was deorbited before the invasion took place. In addition to this, the CIA had a Corona satellite that entered orbit on August 7. It carried two capsules for film return. The first one appeared reassuring; it showed no indications of Soviet preparations for an attack. The second capsule returned photos that clearly showed such  preparations, including massing of troops. By the time that film reached Washington, however, those photos were of historical interest only. The invasion had already taken place. 31
Clearly, the CIA needed real-time space reconnaissance, and its pursuit of this goal would represent one more instance wherein a task originally thought to require astronauts would be accomplished using automated electronics. The true solution would lie in doing away with photographic film, which took time to expose and return. This film would give way to a new electronic microchip called a charge-coupled device. With an image focused onto this chip, it would convert the image into a rapid series of bits. The data, transmitted to the ground, would give the desired real-time photography, and with very high resolution. In addition to this, by freeing reconnaissance satellites from the need to carry and return film, this invention would allow such spacecraft to remain in orbit and to operate for years. 32
The charge-coupled device grew out of the work of two specialists at Bell Labs, William Boyle and George Smith. In 1969, such technology still lay in the future. The view in the Air Force was that the CIA would need piloted spacecraft to produce the real-time photos. The late lamented MOL had represented a possible method, for an onboard photointerpreter might take, develop, and analyze photos on short notice. Now, with MOL in its graveyard, attention turned to the Space Shuttle. It might fly into space, execute a single orbit, and return to its base with film exposed less than an hour earlier.
Because much of the Soviet Union lies above the Arctic Circle, the Air Force was accustomed to placing reconnaissance satellites into polar orbits. It could not do this by firing its boosters from Cape Canaveral; geography dictated that these boosters would fly over populated territory. A launch to the north carried the hazard of impact in the Carolinas; a launch to the south would compromise security if the rocket fell on Cuba. Hence, the Air Force maintained its own space center at Vandenberg AFB, on the California coast. It offered a clear shot to the south, across thousands of miles of open ocean. 33
While a satellite orbit remains fixed in orientation with respect to distant stars, the earth rotates below this orbit. This permitted single reconnaissance missions to photograph much of the Soviet Union. However, it meant that if  a shuttle was to execute a one-orbit mission from Vandenberg, it would return to the latitude of that base after 90 minutes in space only to find that, due to the earth's rotation, this base had moved to the east by 1100 nautical miles. Air Force officials indeed expected to launch the Shuttle from Vandenberg, and they insisted that the Shuttle had to have enough crossrange to cover that distance and return successfully.
The Air Force had other reasons to want once-around missions. Its planners were intrigued by the idea of using the Shuttle to retrieve satellites in orbit. They hoped to snare Soviet spacecraft in such a fashion-and because Moscow might defend such assets by deploying an antisatellite weapon, the Air Force took the view that if the thing was to be done at all, it was best to do it quickly. A once-around mission could snare such a spacecraft and return safely by the time anyone realized it was missing.
In addition to this, NASA and the Air Force shared a concern that a shuttle might have to abort its mission and come down as quickly as possible after launch. This might require "once-around abort," which again would lead to a flight of a single orbit. A once-around abort on a due-east launch from Cape Canaveral would not be too difficult; the craft might land at any of a number of sites within the United States. In the words of NASA's Leroy Day, "If you were making a polar-type launch out of Vandenberg, and you had Max's straight-wing vehicle, there was no place you could go. You'd be in the water when you came back. You've got to go crossrange quite a few hundred miles in order to make land." 34
The Air Force had ample opportunity to emphasize its desire for crossrange by working within the Joint Study Group that Paine and Seamans had set up to seek a mutually-acceptable shuttle design. There were informal discussions as well. George Mueller, who continued to head NASA's OMSF through the whole of 1969, met repeatedly with Air Force representatives at his home in Georgetown, close to downtown Washington. One of his guests was Michael Yarymovych, an Air Force deputy assistant secretary. Another guest, Grant Hansen, was assistant secretary for research and development. He and Mueller also were co-chairmen of the joint study. 35
 These Air Force leaders knew that they held the upper hand. They were well aware that NASA needed a shuttle program and therefore needed both the Air Force's payloads and its political support. The payloads represented a tempting prize, for that service was launching over two hundred reconnaissance missions between 1959 and 1970.36 In addition to this, Air Force support for a shuttle could insulate NASA quite effectively from a charge that the Shuttle was merely a step toward sending astronauts to Mars.
Yet while NASA needed the Air Force, the Air Force did not need NASA. That service was quite content with existing boosters such as the Titan III. "Sure, NASA needs the shuttle for the space station," Hansen said in the spring of 1970. "But for the next 10 years, expendables can handle the Air Force job. We don't consider the Shuttle important enough to set money aside for it."
Yarymovych has a similar recollection:
It was not that simple, of course; no impromptu discussion in Mueller's home would settle such an issue. Rather, it was a matter for the formal protocols of Air Force-NASA cooperation. There was strong conflict between these agencies' wishes, for a NASA baseline document of June 1969, "Desired System Characteristics," emphasized that NASA needed only 250 to 400 n.mi. of crossrange, enough to assure a return to Cape Canaveral at least once every 24 hours. Faget's straight-wing shuttle could achieve 230 n.mi.; straightforward modifications would meet NASA's modest requirements.
To give the Air Force its 1100 n.mi. of crossrange would impose a serious penalty in design, by requiring considerably more thermal protection. The change to a delta wing, even without crossrange, would add considerably to the wing area demanding such protection. Crossrange then would increase this requirement even further. The Shuttle would achieve its....
....crossrange by gliding hypersonically, and hence would compromise the simple nose-high mode of reentry that would turn it into a blunt body. This hypersonic glide would produce more lift and less drag. It also would increase both the rate of heating and the duration of heating. Crossrange thus would call for a double dose of additional thermal protection, resulting in a shuttle that would be heavier-and more costly. 38
Even in its simple straight-wing form, Faget's concept of a two-stage fully-reusable shuttle did not take NASA by storm. It won its pre-eminence only after a process of review and evaluation that extended through 1969 and into 1970. The framework for this process involved the contractors' studies of shuttle configurations that had begun early in 1969. Those studies ruled out expendable boosters for a reusable shuttle, for such boosters were found to exceed the Saturn V in size. Fully- and partially-reusable shuttle concepts  remained in the running, and Faget's concept counted as a new example of the former. NASA proceeded to examine it alongside several alternatives, beginning in mid-1969.
The initial round of studies, during the first half of 1969, had come to $1.2 million, divided equally among four contractors. NASA now extended these studies by giving $150,000 more to each of three contractors, with McDonnell Douglas receiving $225,000. The participating companies also received new instructions that redirected their work.
North American Rockwell had examined expendable boosters. With this approach now out of favor, this firm was free to direct its attention to something new. This proved to be Faget's straight-wing concept, largely in the form he recommended.
McDonnell Douglas, which had examined its Tip Tank stage-and-a-half design, now switched to two-stage fully-reusables. These, however, were not Faget's, but rather continued an earlier line of work. They featured orbiter designs derived from the HL-10 lifting body, with this contractor's engineers considering 13 possible configurations for the complete two-stage vehicle.
At first, the other two contractors saw little change in their assignments. Lockheed was to continue with studies of Star Clipper and of its own version of the Triamese. General Dynamics, home of the initial Triamese concept, was to study variants of this design, and would also apply its background to design a fully-reusable concept having only two elements rather than the three of Triamese.
The orders for this redirection went out on June 20, 1969. Within weeks, the studies brought a flurry of activity that further narrowed the admissible choices. Expendable boosters had already fallen by the wayside. On August 6, a meeting of shuttle managers brought a decision to drop all partially-reusable systems as well. With this, both Lockheed's Star Clipper and McDonnell Douglas's Tip Tank were out. This decision meant that NASA would consider only fully-reusable concepts. 39
Partially-reusable designs had represented an effort to meet economic goals by seeking a shuttle that would cost less to develop than a fully-reusable system, even while imposing higher costs per flight. This approach had held promise prior to the spring of 1969, when the Shuttle had been.....
....considered largely as a means of providing space station logistics. Now its intended uses were broadening to include launches of automated spacecraft, which meant it might fly far more often. The low cost per flight of a fully-reusable now made it attractive, and encouraged NASA to accept its higher development cost. 40
There were at least five ways to build a fully-reusable shuttle, and NASA had appropriate designations and descriptions:
FR-1: the Triamese;
FR-2: a two-stage vehicle with the engines of both stages ignited at launch;
FR-3: a two-stage vehicle with engines in the orbiter ignited only upon staging (Faget's shuttle was an FR-3; so were the concepts of McDonnell Douglas);
FR-4: a variant of the Triamese with the core stage not of the same length as the twin booster stages;
FR-5: a concept designed to avoid a shift in its center of gravity as its propellant tanks would empty, thus easing problems of stability and control.
 On September 4, another meeting eliminated the Triamese configurations. The initial concept, the FR-1, had called for three elements of common length and structural design. It had proven difficult, however, to have one shape serve both as booster and orbiter; to Silveira, "it gets all screwed up, so you get a lousy orbiter and a lousy booster, but you don't get one that does well." Advocates of the Triamese had turned to the FR-4, with its unequal-length design. This, however, proved heavier than the FR-3, while requiring two booster elements rather than one. It also lost much of the potential cost saving from design commonality between the three elements.
The FR-3 and FR-5 remained. The latter had few advocates; the problem of center-of-gravity shift was not so severe as to call for the design innovations of this class of concepts. The manager Max Akridge writes, "It was felt at this time that clearly, the FR-3 configuration was the forerunner." 41
These decisions brought a further redirection in the studies, for while North American Rockwell had gotten an early start on Faget's concept, the other three contractors had to change course. McDonnell Douglas, having found no advantage in its lifting-body orbiters, turned to winged orbiters resembling those of Faget.
While Lockheed also turned to the FR-3, it did not embrace Faget's concept wholeheartedly. This company had spent several years studying Star Clipper, which featured a lifting-body orbiter, triangular in shape. This now looked like a good way to meet Air Force crossrange requirements, and Lockheed's new design retained this lifting body, with a broad underside in the shape of a delta.
General Dynamics showed its own individuality. That firm had designed its Triamese with retractable wings, which would fold into the body during flight but swing outward for landing. This eased the problem of providing these wings with thermal protection, because the fuselage would shield them. This feature now reappeared in the company's new FR-3. It drew on more than Triamese; it also reflected company experience with swinging wings. These were part of the F-111 fighter-bomber, which swung its wings to achieve good performance in both subsonic and supersonic flight.42
 Hence, by the end of 1969 NASA had settled on the FR-3 as its choice, with Faget's specific concept in the forefront. This raised important questions concerning thermal protection. The booster was to be as large as a Boeing 747, yet was to outperform the X-15, reaching considerably higher speeds. The orbiter would be longer than a Boeing 707. For both, the thermal protection had to be reusable.
Within the industry, a standard engineering solution called for the use of hot structures. This approach had a background that included the X-15, Dyna-Soar, ASSET, as well as the Lockheed SR-71 that was flying routinely above Mach 3. Hot structures typically called for titanium as the basic material, covered with high temperature insulation and an outer skin formed of metallic shingles. The metal was molybdenum or columbium, to withstand extreme temperatures while radiating away the heat. Like the shingles on a roof, those on the surface of a hot structure were loosely attached, to expand and contract freely with temperature change.
Such structures were complex, and the shingles posed difficulties of their own. Columbium and molybdenum oxidize readily when hot, and required coatings to resist this. The Dyna-Soar had been designed to use such thermal protection, and Faget declared that "the least little scratch in the coating, the shingle would be destroyed during re-entry." In turn, lost shingles could bring the loss of a vehicle.
NASA and Lockheed now were developing a new surface material: an insulation made of interlaced fibers of silica that could be applied to the outside of a vehicle. These could withstand temperatures of 2500 degrees Fahrenheit, making them suitable for all but the hottest areas on a reentering shuttle. The outer surface would radiate away the heat, in the fashion of the shingles. The thickness of the silica then would prevent most of the heat from reaching the vehicle's skin. This material would not oxidize. It also was light, weighing as little as 15 pounds per cubic foot, or one-fourth the density of water.
This material would form the well-known "tiles" of the Shuttle program, being attached to the skin in the form of numerous small shapes somewhat resembling bricks. In 1969, their immediate prospect lay in simplifying the design of hot structures. These might now dispense with their shingles; engineers instead would use titanium to craft an aircraft structure, with skin  covering an internal framework, then provide thermal protection by covering the skin with the tiles. 43
The design studies of 1969 raised another tantalizing prospect: that these tiles might offer enough heat resistance to build the basic structure of aluminum rather than titanium. Titanium was hard to work with; few machine shops had the necessary expertise. Moreover, its principal uses in aerospace had occurred within classified programs such as the SR-71, which meant that much of the pertinent shop-floor experience itself was classified. This metal could withstand higher temperatures than aluminum. Yet, if tiles could protect aluminum, the use of that metal would open the shuttle to the entire aerospace industry. In Silveira's words, building aluminum airplanes was something that "the industry knew how to do. The industry had, on the floor, standards-things like, 'What are the proper cutting speeds?' They knew how to rivet or machine aluminum." 44
Hot structures, built of titanium, would continue to represent an important approach in shuttle design. As early as 1969, however, Lockheed took the initiative in designing a shuttle orbiter built of aluminum and protected with tiles. General Dynamics added its own concept, featuring aluminum protected by shingled hot structures that could keep internal temperatures below 200 °F.
At the end of 1969, the contractors' orbiter concepts were as follows, with the boosters being similar: 45
North American Rockwell
Faget-type, straight wing
Deployable straight wing
Delta lifting body
Faget-type, straight wing
 These represented variants of Faget's two-stage concept, which showed that shuttle design had come a long way during that year. Twelve months earlier, the candidate configurations included expendable boosters as well as partially- and fully-reusable concepts. The range of alternatives included the Titan III-M with an enlarged Dyna-Soar, the Star Clipper, and the Triamese. These were as mutually dissimilar as a fighter, a bomber, and a commercial airliner.
People still debated such issues as delta wings vs. straight, aluminum vs. titanium, and hot structures vs. tiles. By year's end, however, everyone agreed that the shuttle would look much like Faget's. This meant that the most basic issues of configuration had been settled, allowing engineers to advance to deeper levels of detail. The studies of 1970 would pursue such levels, and would lay important groundwork for the eventual evolution of complete engineering designs, explicit in all particulars.
The work of 1969 had given the space station more support than the Shuttle. The studies of that year initially had allocated $5.8 million for the station and only $1.2 million for the Shuttle; the additional funds granted for the latter at midyear, totalling less than $0.7 million, did little to redress this imbalance. In 1970, however, NASA would bring the shuttle to the forefront. In this year, the centerpiece of effort would involve two $8-million contracts for further work on the shuttle. There would be no significant amount of new funding for the station. These internal NASA decisions would point toward abandonment of the station, at least for a number of years, and elevation of the Shuttle into the sole focus for NASA's future. 46
Station Fades; Shuttle Advances
In preparation for the work of 1970, NASA and its contractors established a new set of institutional arrangements. Following a proposal of Dale Myers, the agency gave responsibility for managing the upcoming study contracts to NASA-Marshall and to the MSC. Each center would hold a complete contract, covering both the booster and orbiter. Marshall, however, would provide technical direction for the booster  portion of both contracts, while MSC would give similar direction for the orbiter portions. This continued the arrangements of Apollo, which had assigned responsibility for the Saturn V to Marshall and for the Apollo moonship to MSC. 47
Within the industry, competing companies made their own arrangements. There would be only two principal new study contracts, but there were many more than two firms eager for the business, and several of them proceeded to form teams. This reflected the Shuttle's two-stage design, for the complete shuttle was likely to be too large a project for a single contractor to handle. Accordingly, Lockheed teamed up with Boeing, with the two companies proposing respectively to handle the orbiter and booster. North American Rockwell joined with General Dynamics, while McDonnell Douglas associated with Martin Marietta. 48
Another and highly important set of decisions extended the scope of Air Force-NASA cooperation. The two agencies had collaborated on a joint study of shuttle requirements as part of the work of the STG; this had led to the issuance of a three volume report in June 1969, classified Secret. That collaboration, however, had merely served the immediate needs of the STG and its supporting studies. Now, in February 1970, Paine and Seamans agreed to set up a permanent coordinating committee, with members to be drawn in equal numbers from each agency. As in the 1969 joint study, there again would be two co-chairmen: Dale Myers and Grant Hansen. Hansen had co-chaired the earlier study, while Myers would replace Mueller, on behalf of NASA. 49
On the matter of crossrange, at least for the moment, they agreed to disagree. Neither agency would seek to impose its will on the other. Rather, each main study contract would conduct two design exercises in parallel: one for an orbiter with crossrange of 200 nautical miles, the second with capability of 1,500 nautical miles. The first was well within the reach of Faget's straight-wing concepts; the second called for more than the Air Force would need. Like a baseball player who swings two bats during warmup and then finds that his single bat feels lighter, the exercise of designing for 1500 nautical  miles would stretch the minds of engineers and make it easier for them to achieve the 1100 nautical miles that represented the real requirement.
From the outset, however, Paine and Seamans could agree on an important point which Paine described as "a payload volume of 15 feet diameter by 60 feet long." Previous studies had considered payload bays not only of this size but of 22 feet diameter by 30 feet long. The new requirement reflected the needs of both agencies. 50
The Air Force needed length, for its reconnaissance satellites amounted to orbiting telescopes, and these had to be long to yield the sharpest images. Moreover, such satellites were growing markedly in length. The Corona spacecraft of the 1960s, each with an attached Agena upper stage, had started at 19 feet and quickly grew to 26. The CIA was now readying a new class of satellites that would win the name of Big Bird. With dimensions of 40 by 10 feet, it represented a backup to MOL, whose length and width had been virtually the same. The next generation of satellites, called Kennan, would keep the 10-foot diameter but would grow in length to 64 feet.
Big Bird was in an advanced state of development in 1970; the first of them flew to orbit atop a Titan III-D in June 1971. Their photos gave a resolution of two feet. Kennan was still in initial studies, for it would introduce long-duration operations and the use of charge-coupled devices rather than film. Though it would not fly until December 1976, its images would show resolution as sharp as six inches.
In 1970, the size of Kennan had not been fixed; indeed, very little about this project had been fixed. It was clear to Air Force planners, however, that they would face an increasing need to launch long satellites. Accordingly, they declared that they would need a length of 60 feet for the shuttle's payload bay.
While NASA did not need so much length, its officials wanted a 15-foot diameter to accommodate modules for a space station. This reflected a new approach to the design of such stations. The studies of the 1960s, including those that Paine had initiated in 1969, had envisioned a space station as a single unit that would fly atop a Saturn I-B or Saturn V. As the prospects for Saturns faded while those of the Shuttle seemed to advance, it appeared prudent to envision a class of stations that could be assembled in space as an  array of cylindrical modules, one module per shuttle flight. A shuttle bay with this diameter would accommodate modules 14 feet across, intermediate between the 10 feet of MOL and the 22 feet of Skylab. 51
In addition, a 15-by-60 foot bay would serve the needs of both agencies by providing room for the space tug and its payloads. Many spacecraft would fly to high orbit, including geosynchronous orbit, and the payload bay had to address such expectations as that future communications satellites would also grow larger. Thus, when Dale Myers asked Grant Hansen to weigh the merits of a reduction to 12 x 40 feet, Hansen replied:
On February 20, NASA officials issued the formal Request for Proposals (RFP) that would lead to awards of the Shuttle study contracts of 1970. In addition to studies of the complete two-stage vehicle, a separate RFP solicited proposals for similarly detailed studies of a main engine. Competing companies had 30 days to respond; the submitted proposals would then go before a Source Evaluation Board that would pick the winners. This board included members from the Air Force. On May 12, it chose the teams headed by  McDonnell Douglas and by North American Rockwell and accepted their bids. These companies now were to proceed with 11-month studies that would carry their shuttle designs to new levels of detail, with each study contract being funded at $8 million.
A year earlier, those same companies had won similar $2.9 million contracts for studies of the space station. This reflected their strength, for they were the only firms to have designed and built both piloted spacecraft and large hydrogen-fueled rocket stages for Apollo. McDonnell had built the Mercury and Gemini spacecraft; Douglas' credits included the S-IVB stage, MOL, and the upcoming Skylab. North American's record included the Apollo spacecraft and the S-II, the second stage of the Saturn V. Other companies had gained strong achievements: General Dynamics's Atlas, Lockheed's Agena stages used with Corona, Martin Marietta's Titan family, Grumman's Apollo lunar module. In their experience, however, McDonnell Douglas and North American Rockwell were in a class by themselves. 53
They had initiated their current space station studies in September 1969. By the following July, these companies had carried their designs to a good level of detail. Their stations would take shape as a cylinder with diameter of 33 feet, suitable for launch by Saturn V, surmounted by an enormous solar array. Jack Heberlig, a space station manager at MSC, described the internal layout as "basically four decks with a cellar and an attic." The cellar and attic would house spacecraft equipment, including storage tanks as well as noisy fans and blowers, with acoustic insulation to keep their noise from disturbing the crew. The four decks would provide room for living and working.
Mockups, built by the contractors, showed how crew members would live in comfort. Each person would have a stateroom resembling a small college dorm room, with a bunk bed, desk and chair, television and communications equipment, and plenty of storage space in drawers and a closet. Two communal lavatories would each provide a shower stall, urinal, and a zero-gravity toilet called a "dry john." The commander's stateroom would feature a personal lavatory, a small conference table, and a computer terminal-which was rare in 1970.
A U-shaped galley, with plenty of countertop space, would include ovens, dishwasher, trash compactor, refrigerator, and storage cabinets. Tables, each with four seats, would stand near a wall of freezer cabinets. Within a recreation area, a similar table would provide room for hobbies and games, with television and movies also being available. A medical room with a treatment table would serve as a dispensary and first-aid station, while also supporting studies of crew members' general health.
Though a Saturn V was to launch this station, its logistics were to depend on use of the Shuttle. Heberlig projected "a minimum of 91 Shuttle flights over  a 10-year period." In July 1970, with the Shuttle/station having survived by the narrowest of margins in both the Senate and House, it was clear that to proceed with simultaneous development of both projects would court disaster. 54
Dale Myers responded by deciding to move ahead with the Shuttle first. This brought a major NASA decision that redirected the ongoing space station studies. While work on 33-foot-diameter versions would continue, the agency now would emphasize investigation of modular versions, which the Shuttle could build in orbit as well as service.
This decision took shape as part of the broader cutbacks that Paine announced in September 1970. Those cutbacks canceled two Apollo missions, freeing their Saturn V rockets for other duty, and Paine noted hopefully that they might find use "in the Skylab, Space Station or other programs where manned operations or a heavy boost capability is required." 55
Early in 1969, a year and a half previously, it had appeared obvious that Paine's station, with its crew of 12, would fly atop a Saturn V. The space station studies of that decade had all assumed use of Saturn-class launch vehicles, with the Shuttle merely as a handmaiden, a logistics vehicle that was poorly defined. By mid-1970, however, the Shuttle and station had reversed positions. The Shuttle now was ready to stand on its own, justified in large part by the work it could do in supporting the Air Force's automated spacecraft. With the Saturn V now representing the rarest and most valuable of commodities, NASA could expect to use it only after the most searching examination of alternatives. These alternatives would particularly include the construction of modular space stations, which could avoid use of Saturn-class boosters entirely.
Now it was the station that retreated to the realm of initial studies. There was no clear understanding, at least in 1970, of how specifically one might proceed with such modular stations, for the work to date had emphasized the design of such stations as large single units. North American Rockwell and McDonnell Douglas carried through an appropriate round of studies during 1971, and issued their reports. Then, with budget cuts continuing to squeeze future goals, NASA largely dropped its plans for such stations pending....
....completion of shuttle development. Not until 1984, with the shuttle operational, would President Reagan finally give the go-ahead for a space station program. 56
After 1970, with the space station off in the wilderness, its supporters could at least take heart that the Shuttle's capacious payload bay had been sized to accommodate its modules. The size of that bay also served a more immediate purpose, for it helped to nail down Air Force support. During the latter months of 1970, events showed that by working to win this support, and by deferring the space station to a much later time, NASA was well on the way to overcoming its opposition in Congress.
The NASA appropriation for FY 1971 had been part of an omnibus spending bill that had included funds for the Department of Housing and Urban  Development. In debate during July, the Senate had added $400 million for the programs that Paine described as "sewers." This led Nixon to veto the entire bill, forcing NASA and the other affected agencies to get by for a time through the temporary funding provided by a continuing resolution in Congress. The Senate again took up the appropriations measure on December 7, and once more Senator Mondale introduced his amendment that sought to delete funding for the Shuttle/station.
This time it went down to defeat by a comfortable margin, 26 to 50. Several senators switched their votes to support NASA, with one of them being John Pastore, chairman of the Senate Appropriations Committee. Following the vote, he stated, "This matter came before the Senate twice; the matter has been decided. It has been decided by the House and it has been decided by the Senate." Mondale continued to introduce similar amendments on subsequent occasions, but found his quest becoming increasingly lonely as they met defeat by even larger margins: 22 to 64 in 1971, 21 to 61 in 1972. But with the Senate vote of December 1970, it was already clear that opposition had collapsed. 57
Also during 1970, the joint NASA-Air Force Space Transportation System Committee emerged as a forum where representatives of the two agencies could hammer out solutions in areas of disagreement. The February 1970 concordat between Paine and Seamans, that established this committee, also stated that shuttle development "will be managed by NASA" and "will be generally unclassified." At meetings of this committee, however, Air Force officials proved quite frank in presenting classified material to support that service's point of view.
At a committee meeting on June 29, the Air Force gave a briefing on the size and weight requirements of DoD payloads. This began by disclosing the size and weight of current payloads, and went on to project the specifications of future payloads, eight to ten years ahead. The presentation also reviewed the history of Air Force launch vehicle payload capabilities and the length of payload fairings.
This briefing supported the Air Force demand for a 60-by-15 foot cargo bay. The length would accommodate both current and future payloads; the  diameter would provide room for a space tug. It would also alleviate design complications associated with the restrictions of current launch vehicles such as the Titan III-D, which limited spacecraft to diameters of 10 feet. 58
NASA willingly accepted this requirement, incorporating it explicitly within the Study Control Document that would guide the new round of space shuttle studies. However, this document did not specify a payload weight. At that same meeting on June 29, the Air Force declared that it wanted the Shuttle to carry 40,000 pounds to low polar orbit. This again would provide room for growth; Big Bird, a year away from its first launch, weighed close to 30,000 pounds, and future spacecraft would certainly be heavier.
NASA officials gave their response at another committee meeting, on October 2. They noted that the baseline mission, described in the Study Control Document, had continued to involve logistics resupply of a space station or space base, in an orbit with an inclination of 55 degrees. The Shuttle was to carry 25,000 pounds to this orbit. Following reentry, however, it would not glide to its home base, but would fly with power from turbojet engines. New studies now showed that an operational shuttle could indeed glide safely to this base, dispensing with those jet engines. That would save weight; the Air Force's requirement of 40,000 pounds then would indeed be achievable. 59
By now it was clear that the Air Force was very much in the pilot's seat when it came to steering the Shuttle program. Max Faget learned this late in 1970, when he wrote a memo to his deputy director at MSC: "The USAF appears not to be nearly as firm on the 15 ft. diameter requirements as they are in length. NASA has no need for 15 ft. diameter either. It is suggested that you attempt to have the payload diameter reduced to 12 ft."
Faget was a power within NASA. He was Director of Engineering and Development at the MSC and reported directly to the head of that center, Robert Gilruth. It took the Air Force only three days to put him in his place, with a reply that read: "The USAF fully supports and stands firm on the present Level I requirement for a payload diameter of 15 feet and a length of 60 feet." This reply came from one Patrick Crotty-whose rank was no higher than major. 60
 Late in December, NASA formally upgraded the Shuttle's status. It had been managed by the Space Shuttle Task Group, headed by Leroy Day. The space station had resided within a similar task group. Now the Shuttle received a separate program office, headed by Charles Donlan, Deputy Associate Administrator for Manned Space Flight. Donlan reported directly to Dale Myers, an Associate Administrator, and also kept Day as his own deputy. On NASA's organization chart, this raised the Shuttle to the status of such programs as Mercury, Gemini, and Skylab.
A month later, on January 19 and 20, 1971, NASA hosted a meeting in Williamsburg, Virginia, that included representatives from shuttle study contractors and from the Air Force. This meeting had the purpose of defining shuttle requirements that would guide the work of these contractors. NASA used the occasion to give the Air Force everything it wanted. In particular, the Shuttle would have a delta wing, with crossrange of 1100 n.mi. Its payload capacity, 40,000 pounds into polar orbit, would correspond to 65,000 pounds in a due-east launch from Cape Canaveral. 61
One sometimes hears that when two parties are in a relationship, the one that wants it more is the weaker. NASA certainly had been pursuing support for the Shuttle with unmaidenly eagerness, and the Williamsburg rules were the result. The agency now was promising to build a bigger and heavier shuttle than it had wanted for its own uses, with considerably more thermal protection. It also was prepared to treat the Shuttle as a national asset-which meant the Air Force would not pay for its development or production and yet would receive the equivalent of exclusive use of one or more of these vehicles, entirely gratis. That service would not receive the Shuttle on a silver platter, but would pay for construction of its own launch facilities at Vandenberg AFB. With the Air Force having by far the larger budget as well as greater political clout, the Williamsburg agreement resembled a treaty between a superpower and a banana republic.
Max Faget, for one, was not about to bow to the Air Force's superior wisdom. He was responsible for providing technical direction on the orbiter portions of the shuttle study contracts. His designers duly proceeded to turn out representative designs of delta-wing configurations that could guide the work of the contractors. His heart, however, remained with the straight wing.  He continued to come forward with new design variants until well into 1971, when as Dale Myers recalls, "I just denied MSC further activities on the straight-wing version." 62
Having carried through its elaborate courtship (some people would prefer a different word) of the Air Force, NASA was now about to reap its reward. This came in March 1971, when Air Force Secretary Seamans presented testimony before the Senate space committee:
While Seamans was not ready to give full consent, or to promise to give up the Air Force's cherished Titan III family, he certainly was saying "maybe," and was saying it emphatically.
The Shuttle program was advancing in another respect. With studies of its main engine having been under way for some time, NASA was about to award a contract for its actual development. With this engine as a long-lead item, this would be the first major component of the Shuttle to reach this level of commitment, advancing beyond the level of studies and design exercises to become a true hardware project.
 The Space Shuttle Main Engine
The legacies of the 1960s include the use of hydrogen as a rocket fuel, which powered two important engines of the period. The RL-10, developed by Pratt & Whitney, had a rated thrust of 15,000 pounds. Two of them were in the Centaur upper stage. The J-2 of Rocketdyne was an important component of Apollo, with five of these engines in the S-II stage of the Saturn V and a single J-2 in the S-IVB. Its thrust was 230,000 pounds.
The RL-10 and J-2 certainly did not represent the last word in rocketry. As early as 1967, well before the Shuttle began to take shape, the Air Force initiated an advanced propulsion program that led to new work at both Rocketdyne and Pratt & Whitney. These firms selected different approaches toward improving the hydrogen-fueled rocket engine, with the intent of building test hardware.
At Rocketdyne, the point of departure involved an inescapable shortcoming of conventional rocket nozzles, which had the shape of a bell. Within these nozzles, during and after liftoff, atmospheric pressure retarded the free expansion and outward flow of an engine's exhaust. This reduced both its thrust and exhaust velocity-and did so just when the launch vehicle was heavy with fuel and was burning propellant at the most rapid rate, thereby needing all the thrust and performance it could get.
The cure appeared to lie in a new type of engine, the aerospike. It required a ring-shaped combustion chamber surrounding a central body that resembled an upside-down volcano, with inward-sloping flanks and a central vent. Turbine exhaust flowed through the vent; the main engine exhaust expanded against the flanks, with no wall or barrier separating this exhaust from the atmosphere. Atmospheric pressure thus worked freely to shape the exhaust plume as it exited the engine. The aerospike concept offered a compact engine installation that could perform nearly as well at sea level as in a vacuum. However, this performance was somewhat less than could be achieved with a conventional bell nozzle. 65
Accordingly, Pratt & Whitney accepted the disadvantages of the bell and sought to achieve the highest possible performance by raising the engine's internal pressure. The J-2 had operated at 763 pounds per square inch. Pratt....
.....& Whitney expected to go much higher. This pressure increased the exhaust velocity. It also allowed an engine to produce high thrust within a compact and lightweight package.
The Air Force awarded a contract to Pratt & Whitney for a new engine, the XLR-129. It aimed at 2740 psi, nearly four times the pressure of the J-2, and was to deliver 250,000 pounds of thrust. NASA went on to support the work at Rocketdyne, which went forward with its own program for the development of aerospikes. Thus, when concepts of the Shuttle began to jell, during 1969, these programs offered two paths toward development of a shuttle main engine. 66
Midway through that year, Wernher von Braun, director of NASA-Marshall, sent telexes to the nation's rocket-building companies that asked a.....
.....number of specific questions. One appreciates the flavor of his queries by noting a few:
Other questions raised searching issues in a host of technical areas: turbine design for high temperatures, high-speed turbopumps, seals and pump bearings, and ground-test facilities, onboard engine checkout using computers. Von Braun also expressed concern that engine materials would become brittle when exposed to hot hydrogen at high pressures. In addition to this, he sought to uncover shortcomings in the aerospike, which faced possible problems of delayed ignition, combustion instability, sources of hot gas to drive the turbines, and the credibility of estimates for component weights and efficiencies. 67
The aerospike held on through the summer, with shuttle managers not only continuing to consider it on an equal basis with the high-pressure bell, but even looking at shuttle designs offering interchangeability between both types of engine. NASA and its contractors, however, had no real experience with the aerospike, though they had plenty with the bell; indeed, all their rocket engines built and flown to date had been of the bell type. In October, a meeting of shuttle managers brought a decision to use the bell type only. This decision won unanimous support from key technical people at both NASA-Marshall and the Manned Spacecraft Center. With this, the aerospike was out in the cold. 68
This was bad news for Rocketdyne, which had conducted a limited amount of high-pressure work but nevertheless had spent several years placing its money on the wrong horse. It was correspondingly good news at Pratt & Whitney, which had a solid head start. That firm had already built and tested a high-pressure thrust chamber, suitable for the XLR-129, though that chamber had lacked its own turbopumps. High-performance turbopumps, however, were becoming a specialty of the house within this company, which was also pursuing a NASA project. It was building versions suitable for a high-pressure engine of 350,000 pounds of thrust. 69
Even so, NASA was not about to accept the XLR-129, for that engine promised 250,000 pounds of thrust and NASA wanted 415,000, to reduce the  number and weight of the engines in the Shuttle's booster. In turn, those 415,000 pounds were to come from an entirely new rocket motor, the Space Shuttle Main Engine (SSME). Still, it was highly likely that this SSME would take shape as an enlarged XLR-129.
In February 1970, NASA issued a Request for Proposal that would lead to the award of three contracts, each funded at $6 million, for detailed engineering studies of the SSME. These contracts went to Rocketdyne, Pratt & Whitney, and Aerojet General. Aerojet had its own strong base of experience; it had built the hydrogen-fueled NERVA test engines in collaboration with Westinghouse, and had also built main and upper-stage engines for the Titan family. Though the Air Force now turned the XLR-129 effort over to NASA, Air Force engineers remained active within this program, transferring XLR-129 technology to Rocketdyne and Aerojet and making sure that its lessons were understood.
During 1970, Pratt & Whitney showed anew that it had the team to beat. In its work on the NASA 350,000-pound engine, it built a hydrogen turbopump that produced over 100 horsepower per pound of weight of its turbomachinery. In World War II, nearly 30 years earlier, builders of aircraft piston engines had counted it as a milestone to achieve a single horsepower per pound; the new turbopump thus was 100 times better. It also had five times the power density of Rocketdyne's J-2, which offered only 20 horsepower per pound.
In August, Pratt & Whitney demonstrated a hydrogen turbopump for the XLR-129. This test drove its turbine with a flow of hot gas from a preburner, a high-pressure auxiliary combustion chamber. This work had particular significance because it went forward unusually quickly. In developing turbopumps for the 350,000-pound engine, this company had taken two years to raise the working pressure of the hydrogen pump to 6000 psi. Its engineers attained 6700 psi with the new XLR-129 pump in no more than six months.
The company also used XLR-129 hardware as a testbed for the SSME. The latter was to have a combustion chamber pressure of 3000 psi. While the XLR-129 had a design pressure of 2740 psi, tests of its thrust chamber repeatedly demonstrated successful operation at and above the level of the SSME. Pratt & Whitney went on to conduct some 200 test firings at these elevated pressures. 70
 At Rocketdyne, the head of the SSME effort was Paul Castenholz, a corporate vice-president who had previously been project manager on the J-2. While he knew he would need more than paper studies to win against Pratt & Whitney, he also saw an opportunity to go that company one better. The complete SSME would include preburners driving turbopumps, a main combustion chamber fed by an injector, and a nozzle. To build SSME-class turbopumps was out of the question; the work would take too long. The rest of the engine was another matter.
Castenholz saw that he could build a complete thrust chamber with everything but the turbopumps. He could craft it using SSME materials and manufacturing processes. Lacking pumps, he would have to feed its propellants using tanks under high pressure. He then could run tests that would demonstrate essential features of a successful SSME: a thrust of 415,000 pounds, stable combustion, a chamber pressure of 3000 psi, and cooling of the engine. He could do these things at full scale. Pratt & Whitney had already done most of them repeatedly, but only in an XLR-129 thrust chamber, of 250,000 pounds of thrust. By leaping ahead into the realm of the SSME, Castenholz hoped to put his firm back in the race.
The recent NASA engine study contract provided no funds for this work. He approached the president of Rocketdyne, William Brennan, and asked for up to $3 million in company money. Brennan sent him on to meet with Robert Anderson, president of the parent company of Rockwell International, who approved this expenditure. Castenholz's engineers then set to work, with an important concern being the prevention of combustion instability within the engine.
Good injector design would suppress this instability. Castenholz started with an injector based on that of the J-2, which had shown its stability during this engine's repeated operational use. He then added technical features that experience had shown would promote even more stability. In the words of a close associate, Robert Biggs, "We put two big preventers on an injector that was basically stable to begin with." While this was like wearing both belt and suspenders, it offered a reasonable guarantee that the thrust chamber would work properly on its first try.
The engine testing proceeded at the Nevada Field Laboratory, a rocket facility some 20 miles northeast of Reno, in the Virginia Mountains. The initial work, late in 1970, involved an uncooled thrust chamber that worked as a....
....heat sink, absorbing heat within thick metal rather than using flows of propellant to carry the heat away. While the thickness of the metal made the engine strong and unlikely to explode, it still could run only very briefly before it would burn a hole in its side and blow up. The first tests dealt only with starting the engine, with ignition trials that ran to durations less than five seconds and pressures that stayed well below rated levels.
Early in 1971, the cooled thrust chamber was ready, aiming at NASA's requirements: 415,000 pounds of thrust, 14,670 ft/sec in exhaust velocity, 3000 psi in pressure. The last test achieved full thrust for only 0.45 seconds. It nevertheless bettered these numbers substantially, delivering 505,700 pounds, 14,990 ft/sec, and 3172 psi. This was twice the rated thrust of the XLR-129, and 60 ft/sec greater in its exhaust velocity. Though small, this improvement was significant. It promised a shuttle payload nearly 2000 pounds heavier than Pratt & Whitney's engine might carry to orbit, along with a saving in the program cost of close to $50 million. 71
 No engine, not even the XLR-129, had yet operated as a complete unit, with turbopumps together with its thrust chamber. Pratt & Whitney soon accumulated a total of 2877 seconds of test operation of its XLR-129 chamber, rated at 250,000 pounds, along with its full-pressure tests of advanced turbopumps for both this engine and the 350,000-pound model. Rocketdyne, in turn, had demonstrated successful starting and stable operation within its SSME thrust chamber, though only briefly. Then, in late January, NASA officials changed the requirements, raising the planned thrust of the SSME from 415,000 to 550,000 pounds. This reflected the growing weight of the Space Shuttle as a launch vehicle, which now was to carry up to 65,000 pounds in payload. Though Rocketdyne had not known that this thrust was to increase, its high-thrust engine test, delivering 505,700 pounds in a chamber built for 415,000, meant that it remained close even to this new requirement. By contrast, the best of Pratt & Whitney's achievements, at 350,000 pounds, lagged well behind. 72
"The highest risk I've ever taken, in terms of a rocket engine, was to build this full-scale thrust chamber for a proposal," Castenholz recalls. "We worked around the clock. We slept at Rocketdyne in the hospital, every night for a month." Though this did not help his marriage (he and his wife Marilyn later divorced), it did provide what he would need to take to NASA.
High-speed cameras had filmed the tests in Nevada. Those films now were ready to show to officials at NASA-Marshall, who were managing work on the SSME. Castenholz arranged to take along the actual thrust chamber: "We thought it was necessary that everyone who would be on the evaluation program should see that we'd actually done it." He wanted them to see his chamber and touch it, not just read about it: "If you can touch something, you feel more comfortable."
Castenholz made a presentation to officials that included Eberhard Rees, the center director. He had succeeded Von Braun, who had taken a position at NASA Headquarters in Washington. Robert Biggs, Castenholz's associate, described this as "the best briefing I've ever seen." It included multiple slide projectors and movies with sound, and not all the scenes were of rockets. When Castenholz said that they had done the test in winter, his slides showed the desert covered with snow. Following the briefing, Rees turned to a colleague and said, "Now I really believe it can be done."
 It was March 1971; NASA now was issuing the Request for Proposal that would lead to the award of a contract for actual development of an operational SSME. Rocketdyne's proposal included an executive summary, seven volumes on technical issues, five volumes on management, and 81 more of supporting data. "Thousands of pages, and beautifully done," Castenholz recalls. "A thick document devoted to every detail. Materials. What the materials included. How it was manufactured-pictures of the manufacturing. The tools we used. The test program-how many tests. Each test. Why. Where."
Copies of the proposal went off to NASA-Marshall on April 21, with each complete copy filling a bookcase. The denouement came on July 13. Castenholz recalls that it was rather casual: "I got a call one day. Bill Brennan called me in and said, 'We won.' Bob Anderson came the next day." Then Rocketdyne held a party, inviting all participants. Castenholz recalls feeling "tremendous, joyous that we'd accomplished it. The idea that we'd won was almost mystical." He thought of things that could have gone wrong, but concluded, "I was fortunate to lead a good team. What won? Running that chamber won, and writing a super proposal." 73
There was no joy in West Palm Beach, Florida, the home of Pratt & Whitney. That firm had taken out advertisements in major aerospace magazines, stating, "We can't wait to start working on the SSME. So we haven't." It had solicited help from the local congressman, Paul Rogers, who had led his state's congressional delegation-including both senators-in writing a letter to Nixon on behalf of this company and its proposal. (Nixon had failed to intervene, and it would not have escaped his attention that California had more electoral votes than Florida.) There was still a possibility, however, that Richard Mulready, Castenholz's counterpart, might yet win the chance to build the SSME. Rocketdyne's margin of victory had been narrow indeed: a score of 711 to 705 in the score of the source evaluation board. It might not take much to tip the balance in the other direction.
Accordingly, the president of Pratt & Whitney, Bruce Torell, lodged a formal protest with the General Accounting Office, with his legal representatives filing a 100-page brief that listed six areas of complaint. The attorneys  asserted that Rocketdyne's proposal held technical deficiencies that violated NASA specifications. NASA also had allegedly failed to conduct meaningful negotiations, had treated the Pratt & Whitney proposal in an "arbitrary and capricious" manner, and had accepted procedures that "maximize the risk of cost overruns." Rocketdyne had supposedly obtained an unfair advantage by diverting funds from the Saturn program to support its SSME effort. A sixth cause of complaint was perhaps the most heartfelt: "Selection of Rocketdyne wastes 11 years of knowledge, test-proven design, and government investment in prior Pratt & Whitney programs." The company had been working on high-performance hydrogen-fueled engines since 1960. Yet if the contract award could not be overturned through this appeal, the firm would have nothing but its RL-10 engine for the Centaur, developed a decade prior to 1971.
This appeal put the contract award on hold. The Comptroller General, Elmer Staats, would not negotiate directly with the competing companies; rather, he would deal with NASA, which had reviewed the contracts and had held the legal responsibility to conduct the competition fairly. NASA's case now harmonized with Rocketdyne's: that the agency had followed proper procedures and had conducted a valid assessment of the proposals. Rocketdyne retained its own legal counsel; Castenholz recalls that "we had an attorney in New York who handled our case." Until Staats made his decision, however, the SSME was moribund. 74
During that summer of 1971, it became clear that the Shuttle program as well was in serious trouble. Not only had it failed to win presidential authorization to proceed; it also was receiving increasingly severe treatment at the hands of critics. While NASA's pact with the Air Force had stilled most of the doubters in Congress, these critics would not be mollified so easily. They were in the Bureau of the Budget.