Most of the preceding chapters touched on ways in which the Defense Department assisted NASA. Units such as the Defense Supply Agency, which administered many NASA contracts, the Army Corps of Engineers, which managed NASA's largest construction projects, and the Air Force, which detailed officers to serve as program managers and directors of center operating divisions, provided essential services in support of the agency. This was in addition to the early, once-only transfers of the Saturn project and other launch vehicles, spacecraft like Tiros, contractor-operated facilities like the Jet Propulsion Laboratory, and the technical skills of the von Braun team. Simply to list examples, however, gives only the barest hint of the significance for NASA of the totality of the support; the Department of Defense (DOD) was the one Federal agency with which NASA had to come to terms in order to carry out its mission at all. The essence of their relationship had far more to do with mutual need than with philosophical arguments concerning the existence or the desirability of one space program or two. The Space Act only outlined the scope of interagency relations in the most general way. The act declared that, while aeronautical and space programs would be managed by a civilian agency, "activities peculiar to or primarily associated with the development of weapons systems . . . or the defense of the United States" would remain DOD's responsibility; and it enjoined NASA to make available "to agencies directly concerned with national defense... discoveries that have military value or significance." 1 It is as well, then, to set aside preconceptions. "Civilian" and "military" are not the same as "peaceful" and "nonpeaceful"; duplication of programs could be "warranted" or "unwarranted"; and much of the struggle over the military uses of space was as much between elements within DOD as between DOD and NASA. In short, the principles underlying the U.S. space program resulted from the many subordinate agreements subsisting between the two agencies. One has to begin from the particular  to arrive at the general because, in this sphere, "no clear-cut divisions of responsibility exist or are possible, and it is difficult to describe the situation in general terms."2
What were the elements comprising the NASA-DOD relationship? In at least four ways their interests impinged on each other: common technologies; NASA's continuation of NACA's support of military aeronautics; NASA's overwhelming dependence in its early years on the launch vehicles and ground support provided by the Air Force and on the Saturn rocket and von Braun team transferred from the Army; and the persistent attempts by the Air Force to investigate the military applications of space, especially of manned Earth-orbital operations. As to common technology, there is no discontinuity between civilian and military R&D, no line that one can arbitrarily draw with rocket guidance and data processing on one side, avionics and solid-state physics on the other. A launch vehicle is only a modified ballistic missile; and it cannot be overstated that for everything between sounding rockets and the Saturn I,* NASA relied on vehicles successfully developed by the Air Force between 1954 and 1959-notably the Atlas, Thor, and Titan ballistic missiles in their original or modified versions. This shared technology also signified that NASA and DOD would have to coordinate programs to avoid "unwarranted" duplication of launch vehicles and facilities. Officials of both agencies would want to exchange information in areas of mutual interest, such as space medicine or bioastronautics, in which NASA and DOD were simultaneously performing thousands of research tasks. Indeed, few areas of NASA's R&D were without potential military application. The Surveyor program-than which nothing could seem more unmilitary-is a case in point. Its retromotor, designed to reduce the spacecraft's approach velocity, was used as the second stage of the Air Force's Burner II launch vehicle; the automatic landing system could be applied to vertical or short take-off and landing (V/STOL) aircraft; while the lightweight, remotely controlled TV camera could be used in military communications.3 Hence the crossovers between military and civilian programs: flying DOD experiments aboard Gemini; transferring NASA's Syncom communications satellite to DOD in 1965; and the use, by Defense agencies, of geodetic and meteorological information supplied by NASA.
The transformation of NACA into NASA did not affect its role in supporting research for the military, except to blur the distinction between support and coordination. It is useful to recall the importance of NACA applied research and the ways by which research concepts were translated into military hardware. The concept of the sweptback wing described in a 1945 report was applied to the B-47 and B-52 bombers and the century series of fighter aircraft. The concept of low-aspect-ratio wings to reduce atmospheric drag at supersonic speeds made possible the F-104 and the X-15 rocket plane. Perhaps the best-known example of the military uses of NACA research was the 1953 paper of Allen and Eggers  that demonstrated the superiority of a blunt nose cone in coping with the extremely high temperatures generated by the atmospheric reentry of a ballistic missile. Similar research continued after NASA's establishment and extended to every area of advanced research and technology, but with this difference: NASA collaborated with the services in refining the concept into the development phase. The X-15, for example, was a joint NACA (NASA)-Air Force-Navy project. Moreover, the same agencies collaborated on solid-fuel propulsion programs, despite the absence of near-term missions in NASA's case. All this was in addition to testing Air Force flight models in the unique facilities available at the Ames and Langley centers; for example, the Langley 18.3-meter vacuum sphere was one of the very few available for "dirty" tests, involving the discharge of contaminating substances such as exhaust gases.
Between 1958 and 1966 almost all NASA launches, other than for sounding rockets, took place at the Air Force Eastern Test Range at Cape Canaveral, Florida, and at the Western Test Range at Vandenburg Air Force Base, near Point Arguello, California.** Even after NASA acquired land just north of the Eastern Test Range in 1961-1962, all NASA launches (except the sounding rocket and Scout satellite launches at Wallops Island) continued on ranges operated by and under conditions stipulated by the Air Force. The nature of NASA programs, however, soon led to disagreements over how NASA would pay DOD for range and tracking support. Moreover, the "single manager" concept of range support ran afoul of the determination of NASA officials to decide their own requirements. It might, for instance, seem like gross and unwarranted duplication when NASA built tracking stations "colocated" near DOD-operated facilities on Antigua and Ascension Island. NASA officials disagreed. The colocation issue could only be decided on the merits of each case, not by assertions of abstract principle. For all the draft and signed agreements prepared by both agencies, the problems of funding and range support were not amenable to any simple or permanent solution, whether a solution involved defining an agency's "unique requirements" or identifying the free. services that each agency provided (or should provide) the other.
For brevity's sake, NASA-DOD relations may be categorized under the headings of support, coordination, and rivalry. The body of this chapter involves a closer look at these categories: how they evolved, the formal mechanisms of interagency relations, and certain problems stemming from the dual management of the U.S. space program. Analyzing specific problem areas should bring us closer to answering the following questions:
The Space Act left the scope of the relations between NASA and the military unchartered. The general principles that evolved were a precipitate of many agreements, programs, and working-level relations. Certain preconditions for working together had to be met. There had to be agreement on which programs properly belonged in either agency, on how NASA would acquire the launch vehicles it needed, and on which officials were authorized to speak on behalf of their agency in relation to the other. Within DOD, there had to be an administrative reshuffling to decide the role of the services in the conduct of military space programs; were they to be concentrated in one, parceled out among all, or located in the Office of the Secretary of Defense?
But although the services might agree on little else, they were as one in opposing the pretensions of the Advanced Research Projects Agency (ARPA) to become the military space agency, since its authority would have been at their expense.4 Less than two years after its establishment, ARPA had become a job shop for research tasks that none of the services happened to be doing. The 1958 reorganization led to the creation of the Directorate of Defense Research and Engineering, whose charter gave its head the authority to "approve, modify, or disapprove programs and projects of the military departments and other DOD agencies.''5 Not until the end of 1959 was the Director's authority over ARPA made explicit, and it was the effect, more than the cause, of the process by which ARPA was divested of its most important programs: the Tiros weather satellite went to NASA; the Transit navigation satellite, to the Navy; the early-warning Midas satellite and the Samos reconnaissance satellite, to the Air Force; and Notus, the interim communications satellite system that became Advent, to the Army. ARPA would be a job shop, responsible only for developing advanced systems and turning them over to one of the services at the point at which each system became operational.
The Army's ill-fated Advent program is worth a closer look because of its relation to NASA's Syncom communications satellite.6 The independent origin of both programs and their ultimate convergence may serve as a case study of features common to civilian and military programs: shared technologies; the risks inherent in advanced R&D, particularly where the launch vehicle, spacecraft, and ground support were proceeding concurrently; the distinction drawn by NASA between developmental and operational phases of R&D; and the problems in setting jurisdictions proper to each agency. Defense Secretary McElroy had  created ARPA in February 1958 with the partial objective of initiating and developing space programs for which there was no clear service jurisdiction. One of these, the communications satellite program, was transferred from ARPA to the Army in September 1959, renamed Advent in February 1960, and effectively terminated in May 1962. From its inception Advent was designed as a synchronous communications satellite-a satellite that orbited at the Earth's speed of rotation at an altitude of 35 900 kilometers and appeared as a fixed point in the sky. Three such satellites can provide worldwide coverage, need fewer ground stations than a medium-altitude satellite, and do not present complicated tracking problems. However, the theoretical advantages of synchronous coverage can be nullified by technical difficulties. A synchronous satellite needs a more powerful launch vehicle than one in medium orbit; it needs more electronic parts and, as a matter of course, a backup satellite in orbit in case of failure; and a spacecraft in synchronous orbit can be jammed more easily than a number of medium-altitude satellites, since, being stationary, it can be pinpointed more readily.
The origins and purpose of Advent were such that they seemed to involve no duplication between DOD and NASA. Defense had a pressing requirement for an operational communications satellite; NASA, as an R&D agency, did not. Under a November 1958 agreement, ARPA was placed in charge of active communications satellites, i.e., spacecraft that receive, amplify, and send radio signals, while NASA would work on passive satellites-balloons, for example-which merely reflect signals sent from ground stations. But this arrangement was not binding. "NASA was not foreclosed from research and development in the active satellite field; in fact the agreement recognized that at an appropriate time, and making full use of the knowledge derived from the military experience, NASA would do some development work on components and prototypes for a nonmilitary communications satellite system."7. This is precisely what happened. As the Army proceeded with Advent, NASA contracted with Hughes Aircraft in 1961 for a lightweight synchronous satellite. Writing to Webb in June 1961, Deputy Secretary of Defense Roswell Gilpatric indicated that he considered NASA's Syncom as "complementary to, not duplicative of, the military Advent."8 By agreement that August, NASA undertook to develop the satellite, with DOD furnishing the ground support. Advent was running into serious difficulties. Not only had the program overrun its original cost estimates, but the Centaur upper stage of the launch vehicle that would place Advent in orbit (which was being developed by NASA) had been delayed for more than three years, and no earlier launch would be possible until the Titan III then being developed became available. Additionally, the triode tube, on which Advent was based, was already obsolete. Under the circumstances the program had to be dropped in favor of a medium-altitude communications satellite. But although such a program was approved, it ran into delays while DOD officials negotiated fruitlessly with the Communications Satellite Corporation (ComSat) to provide services. Almost by default NASA, by 1963, had become the only Federal agency working on a program for a synchronous communications satellite.
 The rest of the story may be summed up briefly. NASA launched and placed in orbit three Syncoms in 1963-1964, the first of which never responded to ground signals, while the second (launched 26 July 1963) and third (19 August 1964) operated successfully. At the end of 1964 the orbiting satellites were transferred from NASA to DOD as forerunners of an Initial Defense Communications Satellite System. In June 1966 a single Titan III inserted seven communications satellites into random, near-synchronous, equatorial orbit, creating the nucleus of a worldwide military communications system.
The history of Advent and Syncom illustrates a great deal of the direction in which NASA-DOD relations were tending in the early 1960s. Note particularly Syncom's status as a coordinated rather than a joint project. NASA designed the spacecraft but had no Syncom stations of its own; the ground support, funded and purchased by DOD, was salvaged from Advent. Syncom was a program whose success depended "on the functioning of separate, co-operating systems."9 It was experimental, and it was not designed to handle bulk traffic. Once the research concept had been proved, NASA was willing to turn over the program to the potential user, the more so as NASA had begun to negotiate contracts for "advanced technological satellites" designed to carry several packages of experiments in communications technology. In both communications and meteorology, NASA had to face the question of where development left off and operations began. In meteorology, NASA ceded the Tiros satellite to the Weather Bureau to meet its and DOD's need for an operational weather satellite, while continuing work on the more advanced Nimbus. In communications, Congress created ComSat in 1962 as a profit-making entity for operating the U.S. portion of a future global communications satellite network. There was no need or justification for NASA to develop and operate a system that would duplicate ComSat's Early Bird program for an "experimental-operational" synchronous satellite.10 The transfer of Syncom to DOD thus served two purposes: It gave DOD access to a technology that it needed, and it removed the suspicion that NASA was developing an operational communications satellite system.
But the most important lesson of the Advent and Syncom programs was that the cancelation of one and the transfer of the other would have been much less likely without the agency reorganizations discussed in the following section. The U.S. space program in 1958-1959 was a welter of projects parceled out among NASA, ARPA, and the services. During the next three years, the Director of Defense Research and Engineering became the official charged with the conduct of Defense R&D; ARPA was downgraded to the role of job shop for the Office of the Secretary of Defense; with minor exceptions, the Air Force was charged with conducting the military space program; while the Defense Communications Agency, previously created in 1960, became in May 1962 the "focal point for continuing integration of the space and ground elements of the communications satellite systems to meet Department of Defense requirements."11 As Director of Defense Research and Engineering, Dr. Harold Brown could decide to cancel Advent because the 1958 act created the position that charged him to approve and  review all major Defense R&D programs. The decision to cancel followed a series of reviews at the level of the Office of the Secretary rather than at the service level. The Project 39 review of March 1961 "to study and review major outstanding problems in the Defense Establishment," was followed by the Advent Program Survey that December, which was chaired by the head of the Office of Electronics in Brown's Directorate.12 There had to be officials in both agencies who could, to paraphrase Matthew Arnold, see these programs steadily and see them whole. It is time to turn to the organizational changes that made such oversight possible.
The military space program moved through three overlapping phases from 1959 to 1961. First, the most promising ARPA projects were turned over to the services; the Air Force was made responsible for ballistic missile development in 1959, for military space development generally in 1961, and for military support of NASA in 1962; and the Office of the Secretary of Defense under Gates and McNamara began to demand that the services justify programs by matching costs and benefits. The directive of 6 March 1961 grew out of recommendations by two of President Kennedy's task forces that DOD activities in general and space programs in particular be centralized. It authorized each military department to conduct preliminary research, but it assigned to the Air Force responsibility for "research, development, test, and engineering of Department of Defense space development programs or projects which are approved hereafter." 13 Since the Air Force was already responsible for more than 90 percent of DOD space development, the real force and substance of McNamara's order consisted in bringing the three services under tighter central control by the Secretary, the Deputy Secretary, and the Director of Defense Research and Engineering. The Air Force could not select the projects that were to be developed. DOD officials were explicit about this point; the aim was to prevent interservice conflicts "by further restricting the independent freedom of action of the three military services . . . by limiting the latitude of the military departments to increase emphasis and funding for various projects."14
The McNamara directive tightened and clarified the ground rules for conducting military R&D in space, without explaining the requirements that would justify military space programs or their relation to NASA's programs. Meanwhile, the Air Force conducted a major reorganization of its own. Since 1954 Gen. Bernard A. Schriever had been in charge of the Air Force ballistic missile program as head, first of the Western Development Division of the Air Research and Development Command (ARDC), then of the Air Force Ballistic Missile Division that replaced it in 1957, and finally of the Air Force Systems Command (AFSC), established on 1 April 1961 by a merger of ARDC and the Air Materiel Command. The 1961 reorganization consolidated the Air Force space and ballistic missile programs, with AFSC assuming "direct responsibility for everything  except the logistical support functions."15 It created two new divisions-Ballistic Systems and Space Systems-which became completely autonomous in October 1962, reporting directly to AFSC Headquarters.*** By a series of related changes, the nonprofit Aerospace Corporation was established in 1960 as a spinoff of TRW/Space Technology Laboratories, which wanted to bid on Air Force hardware contracts; the Aerospace Corporation, on the other hand, was responsible for advanced systems analysis and general technical supervision of Air Force missile and space systems. And last, though not least important, a 1962 directive assigned responsibility to the Air Force for support of NASA.16
Four concurrent developments within NASA strengthened the agency's hand in negotiating with DOD. First, NASA and DOD concluded a series of agreements between 1959 and 1963.17 The agreement of November 1959 laid down the principles by which each agency would reimburse the other for services rendered; and to date it has remained the only general agreement, although it has often been modified.18 The agreement of September 1960 formally established the Aeronautics and Astronautics Coordinating Board (AACB), although it had met for the first time three months earlier. The Board's importance, however, was less in what it accomplished for the time being than in its terms of reference. Where the moribund Civilian-Military Liaison Committee created by the Space Act had no authority to implement its decisions, the members of the AACB and its six panels **** were authorized to take actions "utilizing the authority vested in them by their respective agencies."19 The cochairmen were the NASA Deputy Administrator and the Director of Defense Research and Engineering.
These agreements were a sort of organic law for the space program. The Webb-Gilpatric agreement of 23 February 1961 stipulated that neither agency would initiate the development of a new launch vehicle without first seeking the consent of the other.20 The agreement of 14 January 1963 provided that the Merritt Island Launch Area would be operated as a NASA installation, separate and distinct from the Eastern Test Range, although the Air Force would continue to provide common services for NASA and DOD users.21 This superseded an earlier agreement under which Merritt Island would have been operated jointly, and it suggests how far NASA was prepared to go in asserting the right to determine its own needs. One week later, the Gemini agreement confirmed NASA's hard-won independence.22 Here, too, Webb turned down McNamara's proposal of a joint program. The agreement of 21 January affirmed NASA's role as program manager, while stipulating that DOD would participate under arrangements to be made by an advisory Gemini Program Planning Board. Of these five agreements, the first two established the organizational ground rules, the third served as the basic document in coordinating military and civilian programs, while  the fourth and fifth confirmed that NASA was the dominant partner in the manned exploration of space.
These agreements were by no means the only examples of NASA's assertiveness. In March 1962, for example, NASA established independent field installations at the launch facilities of the Eastern and Western Test Ranges. Similarly, NASA sought to participate more closely in the development of vehicles like the Agena upper stage, which it needed for its own launches. By a September 1963 agreement between NASA and the Air Force Systems Command, NASA was assigned a role in launch vehicle planning and became a member of the Air Force Configuration Control Board for the Atlas, Thor, and Agena vehicles.
The second development to affect the conduct of the national space program was President Kennedy's decision to assign the lunar landing mission to NASA. After May 1961 the Air Force had no clearly defined manned mission in space. Its Man in Space Soonest project, when transferred to NASA, evolved into Project Mercury-a nonmilitary project that could not have succeeded without military support. The Air Force Systems Command's 1962 proposal for a "Blue Gemini" went nowhere for lack of support by Air Force Headquarters and by the Office of the Secretary of Defense. When Webb and McNamara signed the Gemini agreement, the only approved manned military program was Dyna-Soar, a winged orbital glider first conceived in 1957-1958 and formally initiated as a program in 1960, when Boeing was chosen to design the glider and the Martin Company to supply the Titan I booster.23 Even before the Gemini agreement, Dyna-Soar had run into trouble; and it is not too much to say that the agreement hastened Dyna-Soar's demise, the more so because the former could do most of what the latter was intended to do, and more. Thus by early 1963 proponents of a manned military program were caught in a merry-go-round of requirements. To justify such a program, they had to explain what a manned mission could accomplish that an unmanned satellite could not; they had to show how a military project could be designed to avoid duplicating what Gemini was accomplishing in Earth-orbital operations; and they found it useful to assert that the Soviet Union had gone further than the U.S. in the military use of space. But Gemini, by including more than a dozen military experiments, inevitably raised the spectre of unwarranted duplication should the Air Force proceed with a manned program.
A third development in NASA-DOD relations was signaled by the establishment of a NASA Office of Defense Affairs in December 1962.***** As with AACB, its establishment was more notable as an expression of policy than for any immediate accomplishment. The new office, headed by retired Adm. W. Fred Boone, was to promote interagency cooperation, serve as "the focal point for all major defense-related matters within NASA, [and] speak for NASA within the framework of established policy."24 NASA and DOD were already linked in many ways: through AACB and its panels; through the Air Force Systems Command liaison office, located in the same downtown Washington office building as the  Office of Manned Space Flight; and through the military detailees stationed at the centers. Treading carefully, for there were many bureaucratic feet to step on, Boone showed discretion in adopting the operating principle that "we would not attempt to have all communications and interactions with the Defense Department channelled through our Office.... we would avoid permitting the Office . . . to become a bottleneck impeding the flow of information and the conduct of NASA-DOD business."25
Fourth, NASA moved to centralize its ground support network. By 1965 NASA had three networks: STADAN (Satellite Tracking and Data Acquisition Network), successor to the Minitrack Network developed for Vanguard and used for tracking unmanned, Earth-orbital spacecraft; the Deep Space Network managed by JPL, which tracked lunar and planetary probes; and the Manned Space Flight Network (MSFN), which consisted in 1965 of nineteen land-based sites plus instrumented ships and aircraft managed by DOD. These networks were complementary in certain respects. MSFN, for example, had to work with spacecraft up to lunar distances and, like the Deep Space Network, it had a number of 26-meter radio antennas for tracking. MSFN and STADAN were sufficiently similar that NASA consolidated them into a single Space Tracking and Data Network in mid-1971, although NASA management never took the final step of completely centralizing the range structure, which would have been undesirable for two reasons. The program offices and their center representatives would have been most unwilling to surrender control of mission operations and tracking, and a persuasive case could be made that "to divorce support completely from direct program control or access is to endanger the responsiveness of that support base to the programs it should serve . . . separately managed support resources by a centralized office may sometimes be more costly than program control since some equipment should be used only during the lifetime of a particular program."26
What headquarters could do, it did. The creation of a unified Launch Operations Center at the Cape in May 1963 removed Marshall management from the scene. The new center, renamed the Kennedy Space Center (KSC) that December, took over Houston's Florida Operations in 1964, and in October 1965 assumed control of all unmanned launches, except for sounding rockets, on either coast. In like manner, the Office of Tracking and Data Acquisition (OTDA) assumed management of JPL's Space Flight Operations Facility, once its prospective use by Pioneer and Lunar Orbiter made central control necessary. As recognition of its coequal status, OTDA was designated a program office in December 1965.
Although these moves provided tighter control over launch support and tracking and data acquisition, headquarters continued to manage the two separately. The furthest that NASA was prepared to go in coordinating them was to create the position of Mission Operations Director with overall responsibility for manned spaceflight programs and to establish an Operations Support Requirements Office (OSRO) staffed by representatives of the program offices, OTDA, and the centers participating in manned spaceflight. Reporting to the Mission  Operations Director, OSRO was charged with considering operations support as a whole, particularly in cases where more than one program and more than one center were involved.27 Equally important, the office was intended to be a single point of contact with the Air Force National Range Division for all manned flight support requirements.28 Whether complete centralization might have accomplished more is debatable. What is evident is the rationalization of NASA's ground support activities and the creation of well-defined patterns of cooperation and support between NASA and DOD. The following section considers these patterns.
While reserving to NASA control of Merritt Island, the Webb-McNamara agreement of January 1963 left open the question of how NASA would pay for those services that the Air Force continued to provide. In 1963 much of what would become the Kennedy Space Center was still a swamp, more a wildlife sanctuary than a launch facility. When the first Saturn IB flew, in February 1966, it was from the Eastern Test Range, as were all subsequent Saturn IB launches for the main Apollo program.****** Only those launches requiring the Saturn V were from KSC's launch complex 39, which was developed by NASA specifically to accommodate them. So heavily did NASA rely on Air Force facilities that the Air Force transferred several launch complexes-launch complexes 34 and 37 for Saturn, launch complex 12 for the Atlas-Agena, launch complex 16 for the Titan II, launch complex 36 for the Centaur-to NASA, since they were already used almost exclusively for NASA launches. Indeed, DOD supported NASA to such an extent that launches of the Navy's Polaris missile at the Eastern Test Range had to wait for those of NASA. 29
The relationship between NASA and the Air Force gave rise to certain issues, in themselves highly technical, yet important judged by the time spent by officials of both agencies in trying to resolve them. With the 1963 agreement NASA became responsible for master planning and developing facilities at Merritt Island, as well as for the preparation, checkout, and launch of its own flight missions. This agreement left open the funding of services that DOD continued to supply, such as the use of launch pads, downrange tracking and data acquisition, and relay, instrumentation, search and recovery ships and aircraft in the Atlantic and Pacific Oceans. The differences over funding were more than merely, technical; they involved each agency's conception of its role in the national space program. Precisely because differences went so deep, none of the interagency, working groups set up to resolve the problem arrived at a settlement acceptable to both sides: not the panel established to work out the details of the Webb McNamara agreement, nor the interagency task forces that studied the problem  from 1965 to 1967, nor the Director of the Budget, into whose lap both agencies finally tossed the problem.30 Basically, the task forces identified four alternative methods of funding. NASA could pay the actual cost of services provided; it could enjoy a host-tenant relationship, paying only for "abnormal" services; it could pay on a pro rata basis, according to some arbitrary estimate of the proportion of the total workload for which NASA was responsible; or it could continue the existing arrangement, under which NASA reimbursed the Air Force only for those requirements that were unique to a given flight project.
In practice, the four options narrowed to two: cost sharing, as DOD preferred, or reimbursement for an agreed proportion of total costs. Why, then, was the problem so intractable? The reasons are complex, intertwined with political considerations that were not always made explicit. First, there was the difficulty of separating the costs peculiar to NASA programs from the total cost of running the Eastern Test Range as a national range used by several agencies. NASA was prepared to pay for certain readily identifiable additional services incurred by the Air Force as part of the NASA workload. The agency accepted the concept that a national range-that is, a facility used by more than one agency-should be run by a single manager. But NASA officials took the argument a step further. Where the range was being operated on behalf of several users, it was impractical to charge each agency on a job order basis.31 What counted was the creation and maintenance of a national resource. In the eyes of agency spokesmen, NASA was consistent, since the same principle had been applied by NASA laboratories to work done for other agencies, including DOD.
NASA officials hinted that no other arrangement could be accepted. Or rather, they intimated that if McNamara insisted on cost-sharing at the Cape, then "NASA would feel obligated to exercise an authoritative voice in planning and management commensurate with its share of funding.33
The history of the funding controversy does have-if not a lighter side-then a kind of awful predictability familiar to students of bureaucratic strategies. Above all, both sides seemed to shrink from any sort of final decision, even to canceling the outdated 1959 agreement. The task force report of December 1963, which recommended that existing funding arrangements be continued, only served as a temporary remedy for confusion. In October 1965 Defense officials reopened the matter by suggesting that NASA cooperate in arriving at an equitable division of costs. This touched off another round of studies and reports that lasted over two years and left both sides as far apart as ever. There were joint working groups,  exchanges of letters, and a budget request by McNamara that struck out $45 million from the Defense budget in support of Apollo-an amount promptly restored by the Bureau of the Budget. By February 1967 NASA and DOD were still so far apart that Boone, who handled negotiations for NASA, and his DOD counterpart "could not even agree as to how we should report that we disagreed."34
These protracted and exhausting negotiations did lead to an interim settlement, however inconclusive. In April 1967 both agencies wrote to Budge' Director Charles Schultze, requesting a judgment on issues pending. Schultze agreed to arbitrate, provided that his decision was accepted by both agencies as binding. When the Bureau's judgment was conveyed to NASA in February 1968 its conclusions left the status quo virtually intact. The Bureau's decision applied to FY 1969 only; and no more final judgment would be possible, pending a financial management system at the Eastern Test Range that would "provide a basis for full identification of costs based upon valid accounting procedures."35 In essence, the Bureau accepted NASA's position on funding; for 1969 NASA and DOD were to pay for range operations on a 60:40 basis, except that NASA would pay 85 percent of the cost of Apollo aircraft. Out of a total Eastern Test Range budget of $260.9 million, NASA would provide $51.4 million, DOD the rest. Such was the inconclusive conclusion to a controversy begun six years earlier.
While the issue of Eastern Test Range funding laid bare fundamental differences over the management of the space program, other funding agreements presented few problems. Where NASA procured launch vehicles from the Air Force, it paid only for the production item, not for development costs. Similarly, when NASA, in April 1964, delegated responsibility to the Defense Supply Agency to administer its contracts, it was agreed that NASA would pay for direct and indirect costs on a per-hour basis. Whatever technical problems might arise, there were no serious philosophical differences comparable to those at the Eastern Test Range. These differences arose in three cases: when NASA believed that DOD support might involve technical direction of NASA programs, when DOD pressed for the consolidation of colocated NASA facilities to avoid unwarranted duplication, and when DOD officials substituted their judgment for NASA's in determining NASA's requirements. The issue of the operation of tracking stations illustrates all three cases. The agencies' positions were lucidly stated by Boone:
NASA, on the other hand,
 NASA officials asserted that the totality of Apollo mission operations had to be under direct agency control; and they drew a line between routine housekeeping services, which they were prepared to delegate, and the technical operation of facilities, which they were not. Moreover, they insisted that colocation and duplication were not the same thing. The NASA tracking stations on Antigua and Ascension Island did not duplicate nearby DOD facilities. The NASA facilities, which operated on the unified S-band system, were part of the Deep Space Network and had nothing in common with the DOD system; they could not be operated by part-time labor pulled from a general-purpose technical pool; and it would serve no purpose and save no money to take the stations out of an integrated mission support system to meet the theoretical requirement of single management.37 It is conceivable that, up to 1962, both agencies might have arrived at a system with similar bandwidth capacity. By 1965 NASA and DOD had gone their separate ways; thus single management remained a very live issue.38
In practice, NASA management was more flexible than a flat assertion of the agency's position would imply. By an agreement of 22 May 1965, NASA and DOD agreed to three principles in the management of colocated tracking stations:39 (1) where facilities were colocated, one agency would be responsible for all base support functions, such as public works, utilities, and logistics; (2) the single manager for base support would normally operate the instrumentation; and (3) where an exception to the above promised cost savings, the agencies would carry out a joint study to devise other arrangements. Thus NASA continued to operate the unified S-band instruments at Antigua and Ascension, although DOD managed both stations.
Another instance in which NASA and DOD had to bend principles pertained to outfitting and managing tracking ships for Apollo. Here the issues were even more complex, since they involved the Navy's Bureau of Ships as well as NASA and the Air Force. In its 1964 budget request, NASA asked Congress for 590 million to convert and equip three ships from the Maritime Reserve for tracking and data acquisition in support of Apollo. When it was learned that both NASA and DOD were planning additional ships without much regard to what the other was doing, the authorization committees withheld funds pending a joint study to resolve several issues. Who would design and operate the ships? Would NASA or the Air Force determine Apollo support requirements? Would the ships be manned by civil service crews or by union crews supplied by Pan American World Airways, the support contractor for the entire Eastern Test Range? And who would operate and maintain the shipboard Apollo instrumentation-NASA or DOD ?
After the usual lengthy negotiations-this time including an ad hoc AACB Committee on Instrumentation-on 15 January 1964 the two agencies signed and sent their agreement to Congress.40 NASA and DOD would work out requirements, with the Air Force placed in charge of centralized planning and management; the Navy would acquire and modify the ships through a specially created project office; and the Navy Military Sea Transportation Service would operate  and maintain the ships at sea. Except in two respects, the agreement gave NASA most of what it wanted. NASA had insisted on ships manned by civil service crews to avoid the possibility of a strike delaying an Apollo mission. The Air Force, however, continued to use union crews as long as its Pan American contract remained in effect; ultimately, they were phased out and replaced by civil service crews. In addition, NASA had demanded a free hand in operating the ships' Apollo instrumentation; as part of the 1964 agreement, NASA relinquished operations to DOD, despite its oft-stated goal of "a fully integrated network operation."41 Yet in virtually every other respect, NASA got what it wanted. The ships were to be operated by the Navy, not the Air Force; NASA, as well as the Air Force, would develop the plans for operating ships "as Range facilities in the Apollo network"; the funds for the ships would be kept by NASA, rather than turned over completely to the Air Force; most significant, "DOD would accept as top priority the Apollo support requirements as stated by NASA. '' 42 In practice, NASA settled for something less than the "totally integrated network" it had considered indispensable.
Launch Vehicle Planning
One of the most critical areas for coordination pertained to the national launch vehicle program. Coordination studies by NASA and DOD went beyond the issue of avoiding wasteful duplication. The development of new booster stages and the phase out of older ones were involved in the long-range planning of both agencies, since the diameter of the vehicle, total thrust, and propulsion system determined the payloads to be launched and the orbits in which they were placed.
Three things should be kept in mind in order to understand the NASA position on launch vehicle development. First, NASA began to develop its largest launch vehicles well in advance of a specific mission requirement. This was true of the giant F-1 and J-2 engines that powered all three stages of the Saturn V and Webb used those as examples of the value of research independent of specific missions.43
Second, certain concepts relating to the national launch vehicle program had come to be widely accepted by 1963: that neither agency should start a new launch vehicle development without joint review; that the number of engine systems should be limited and standardized; that both agencies should cooperate in developing "building block" components that they could both use; and that, in the interest of economy, NASA and DOD should study the possibility of reusable as well as expendable boosters. Several of these concepts had been recommended, or at least discussed, in two influential reports of 1961-1962. The first, prepared by an Ad Hoc Booster Panel of the President's Science Advisory Committee (PSAC) was released early in June 1961, just after President Kennedy went before Congress to propose an expanded national space program.44 The other was the report of the Large Launch Vehicle Planning Group (LLVPG) created by AACB in July 1961, which submitted its final conclusions in February 1962. LLVPG was  to consider the combination of launch vehicles that was best fitted to serve the ends of the manned lunar landing, manned scientific missions, and advanced military operations.45 Neither task force arrived at specific conclusions on the course that the national space program should take. The members of LLVPG disagreed on the mission mode for the lunar landing, hence, on the launch vehicle for Apollo.46 Yet both reports were straws in the wind. There was strong sentiment in the scientific and engineering communities for a thorough investigation of the civilian and military uses of solid- and liquid-fuel propulsion. There was also, by 1963, a livelier awareness of the difficulties in marking out the boundaries of each agency's jurisdiction. As shown, the early understanding that DOD would work on active, and NASA on passive, communications satellites quickly evaporated. The same held for the development of big space boosters, originally NASA's responsibility. By 1961 the Air Force was at work on a large launch vehicle, which would meet its requirements for sending large payloads into Earth orbit and would also be available to NASA. This was the Titan III.47
Third, the national stable of launch vehicles in the early 1970s was very different from what it had been a dozen years earlier. Then NASA had had available seven vehicles, six of them derived from missile programs.48 By 1972 the process of winnowing and sifting and standardizing had gone very far. The only remaining Thor-based vehicle was the Delta, which, along with the Atlas-Centaur, was used for medium and large payloads. The Titan III-C (or the Titan III with the Centaur upper stage) was used for very large packages. The Saturn booster was no longer available, although the remaining vehicles were used to launch Skylab in 1973 and Apollo-Soyuz in 1975. Rejected by the Air Force in favor of the Titan III, the Saturn could have been retained only if Congress had approved an ambitious manned program beyond Apollo. Congress gave no such unqualified approval; NASA suspended production of the Saturn V in 1967 and discontinued it in 1970. Thus the most significant addition to the national launch vehicle program was the use of the Titan III for the largest payloads. The Titan III had several advantages over other large boosters. It could launch multiple payloads up to synchronous orbit; accept a variety of upper stages, including the Centaur, as well as solid-fuel strap-on boosters; and it was capable of using "standardized major components which can be put together in various combinations to perform . . . effectively for different orbits and payload weights."49 Even before the first launch of the Titan III-C, on 18 June 1965, DOD was urging on NASA its feasibility as the booster for Surveyor and Voyager.50 NASA engineers conceded its effectiveness, but they could not use it in place of Saturn. With the completion of the Apollo and Skylab programs, however, there was no further role for Saturn. Of necessity, NASA then turned to the Titan III for the largest payloads.
At no one point did NASA and DOD freeze the number and configuration of their launch vehicles. Instead, they coordinated research and development at several levels. One such way was to narrow the options. The Nova vehicle was dropped when it became apparent that the Saturn would be large enough for the  Apollo mission. Another way was to coordinate advanced research for vehicles for which there was no immediate need. Take, for instance, the history of the large solid-fuel propulsion program.51 In his 25 May 1961 message to Congress, Kennedy designated DOD as the agency to develop solid propellant motors. Although NASA, for technical reasons, chose to concentrate on liquid fuels, it was recognized that NASA's requirement for a solid-fuel motor might be greater than that of the Air Force.+ In December 1963 NASA agreed to fund the program and on 1 March 1965 assumed full management responsibility for developing the 6.6-meter motor, even though no money was provided for it in the 1966 budget request. Webb was called on by several congressional committees to explain why NASA was discontinuing a program developed over such a long period and at such great expense. The provoking aspect of his testimony was his steadfast refusal to say that the program was being "canceled" or "terminated"-only that no more funds would be requested.52 He also contrived to convey that NASA had to do this kind of advanced research in order to anticipate whatever the space program would require. In fact, NASA kept the program alive for another two years through reprogrammings and the partial restoration of funds by Congress. The success of the large-solid-motor program was measured less by the production of a new vehicle than by the demonstration of its feasibility. Nevertheless, such a program could hardly be continued at a time when NASA flight programs were being cut or eliminated from the budget.
Also important in developing a national stable of launch vehicles was the joint study begun in July and completed in December 1964 by the AACB Launch Vehicle Panel.53 The study confirmed that the existing vehicles were adequate to the needs of NASA and DOD. The panel was invited to consider three options, exclusive of the Saturn V and the Scout: continued use of existing vehicles, including the Titan III; the first option less the Titan booster; and the first option less the Saturn IB. The panel concluded that there would be a cost difference of less than 1 percent among these options and that no major savings would result from shifting from present launch vehicles to a system based largely on the Titan III. Coming near the midpoint of the Apollo program, the study confirmed that coordination between NASA and DOD was successful; that the Titan III would be a favored launch vehicle for major programs to come in the late 1960s; and, most significant, that the process of choosing a launch vehicle depended on the mission that the user agency had in mind. The panel examined all the costs, direct and indirect, first of developing, then of producing a launch vehicle; and one of its conclusions was that the production cost was often the least important element in determining whether or not a vehicle should be used. In certain cases, booster costs were as little as 8 percent of total mission costs; the rest was accounted for by the cost of the spacecraft, the integration of the spacecraft and the booster, and the preparation of the mission.54 In sum, the 1964 launch vehicle  study derived more precise estimates for the cost of producing launch vehicles, for substituting one vehicle for another, and for confirming the NASA decision to use the Saturn IB for Apollo.
Interagency Support and Coordination
One can only skim the complex of relations between NASA and DOD. For convenience, "support" and "coordination" can be distinguished as follows: support is where one agency assists in carrying out the other's programs; coordination is where the agencies exchange information on programs being managed by either or both, in order to avoid unnecessary duplication. These definitions are scarcely watertight. In particular cases where the Air Force supported NASA, there was also an element of coordination, since NASA and the Air Force used different contractors for common use items, even in cases where they used the same items.++ Also certain programs, like Gemini, took on some aspects of a joint program as they proceeded. Thus support and coordination are relative terms, and there is a spectrum of possibilities from single management at one end to joint projects at the other.
One of AACB's principal functions was to detect and eliminate wasteful duplication. As one may have come to expect, there is nothing self-evident about the term. The mere fact that both NASA and DOD were sponsoring research in space medicine or microbiology hardly meant that half of it was unnecessary. A 1963 joint review of space medicine in NASA and DOD recognized that duplication was warranted, "if the research problems demanded more than one approach, if there was a scarcity of support in the particular problem area, or if it was necessary for maintaining 'in-house' capability."55 Or consider facilities coordination between NASA and DOD. At the annual AACB review NASA was represented by William Fleming of the Office of Programming, who was succeeded in 1968 by Robert Curtin, the Director of the Office of Facilities. As with basic research tasks or colocated tracking stations, there were no hard-and-fast criteria regarding warranted duplication, but there were certain rules of thumb. A facility that went beyond existing capabilities was not considered as duplication. But even where one facility was similar to another, there were often extenuating circumstances. Duplication could be tolerated where one agency could not handle the other's workload,+++ or where one laboratory was intended to support a program peculiar to one agency,++++ or where one agency considered the other's research to be  in the national interest.+++++ In general, the annual facilities review worked by a process of "distillation"; thus in July 1968 NASA initially considered some eighty projects and ended by reviewing only fourteen.56 The final joint review was the end of a long series of prior decisions. Indeed, some projects were dropped, less because of duplication than for other reasons, for example, because the facility was too far ahead of current needs. In sum, the existence or necessity of duplicating projects or facilities created problems for NASA and DOD that had to be resolved on a case-by-case basis.
These instances of NASA-DOD coordination should suffice to demonstrate the general principles. Most of the studies involving AACB panels were technical and noncontroversial. Some, like the 1968-1969 economy studies, were intended to consolidate base support at the Eastern Test Range and the Kennedy Space Center. Others, like the 1963 review of space medicine research, were information exchanges designed to provide a common data bank. In yet another category were joint studies of research that both NASA and DOD thought potentially useful: a 1966 study of reusable launch vehicles; a 1968 study of the Data Relay Satellite, which could assume many of the tracking and command functions performed by ground stations; or a 1969 study of the feasibility of a single space transportation system for NASA's and DOD's use.57 These examples may stand for a host of others. In very few cases did these studies raise issues that could not be resolved at the working level. Where the two agencies could not agree was in the sphere where program philosophy and program management overlapped, particularly in the cases of Gemini and the Manned Orbiting Laboratory.
Finally, something should be said about NASA support for DOD. The support took three forms: testing DOD prototypes at NASA facilities, sharing knowledge gained in NASA programs with other agencies, and conducting research on behalf of DOD. To exemplify the first category, NASA tested 130 flight models between 1961 and 1969 in Langley wind tunnels.58 This support proved so valuable that officials indicated in 1969 that "DOD expects to ask for help on each new system, in the future placing even heavier demands on NASA." 59 In the second category were the various experiments performed aboard Mercury, Gemini, and Apollo; one might add communications satellites like Syncom, which began as NASA projects and ended up transferred to DOD. In the third category was NASA support for the B-70 supersonic plane and for the "limited warfare" research project on behalf of the services in Vietnam. The Air Force had intended the B-70 to be a manned strategic bomber, despite grave doubts about its feasibility or necessity. By late 1962 it was known in Air Force circles that McNamara would never allow the B-70 to enter production with a new generation of manned bombers; and the plane, renamed the XB-70, became an experimental model that flew instrumentation installed by NASA.. In March 1967 the Air Force transferred management to NASA, after which the XB-70 continued as a joint project for another two years.60 The agreement was advantageous to both sides,  since the Air Force could cut the losses it incurred in sponsoring the prototype of a strategic bomber that would never be built. NASA, on the other hand, wanted the vehicle for its own research, particularly in connection with the civilian supersonic transport program. This case was the reverse of Syncom: What began as a program to develop an Air Force weapon system ended as a joint research project mostly funded by NASA.
A more unusual kind of technical support was provided by the NASA Limited Warfare Committee, established in December 1965 after the Air Force Systems Command asked NASA to develop certain kinds of hardware for use in Vietnam.61 Given the nature of the terrain and the strategy of guerrilla warfare, the services needed unconventional technical support for a war that was not being fought by conventional rules. Most of the research was carried out at Ames and JPL, involved about 100 scientists and engineers, and cost NASA between $4 and $5 million a year. The limited warfare program led to some ingeniously designed equipment: a quick method of patching holes in inflated life rafts, a better parachute steering mechanism, a helicopter that did not make a chopper noise, an acoustic detector that located mortars by ground vibrations, and an aircraft target marker. Although the specific tasks that NASA performed were classified, the fact that NASA was doing this kind of work for the military was generally known. Nevertheless, this support-although it was fundamentally no different from any other NASA research applied by DOD-placed NASA on somewhat shaky ground. Except for the clause in the Space Act that enjoined NASA to make available to DOD "discoveries that have military value or significance," the program lacked clear authorization and made NASA vulnerable to congressional inquiry. For these reasons and because it had achieved its purpose, the limited warfare program was phased out in 1969.
An account of NASA's Gemini program and of DOD's Manned Orbiting Laboratory (MOL) brings into sharper focus all the elements of support, coordination, and rivalry discussed so far. Gemini was a NASA program coordinated with DOD. Under the terms of the Gemini agreement, DOD took part in planning experiments, launch operations, and flight operations. Yet the fortnight preceding the agreement was marked by McNamara's attempt to take over the program, first, by informally proposing a merger of NASA-DOD manned space programs under DOD management, then by formally proposing a joint program.++++++ 62 NASA could not accept either arrangement without compromising its independence. Gemini had been planned to meet the needs of NASA's manned program; it had been under way since December 1961; it had been designed with military needs in mind; and it could not be transferred from NASA without  causing substantial delay. There were also compelling political reasons for Gemini to remain within NASA, since the arrangement by which NASA operated tracking stations in Mexico, Nigeria, Zanzibar, and Spain prohibited their use to support military programs. Under McNamara's terms these stations would be unavailable, and the existence of a joint civilian-military program would jeopardize negotiations for tracking stations elsewhere.63 The effect of transferring Gemini to DOD would be to place in doubt NASA's image as a civilian agency dedicated to the peaceful exploration of space. But beyond these objections, cogent as they were, NASA officials sensed that they could not accede to such a proposal and still retain control over their programs. Just before the meeting that preceded the final agreement, Webb wrote to Seamans that
The Gemini agreement deferred for almost a year the issue of how a military space program might impinge on NASA. Under the terms of the pact, "Gemini was not to be thought of as a joint program, but rather as a program serving common needs, with the Department of Defense paying for the military features, NASA in full charge of the program, and the role of the [Gemini Program Planning] board strictly advisory." 65 The disagreement over Gemini had nothing to do with whether NASA would support the military or not. That support was never in question, only NASA's right to determine how it would provide it. Webb insisted-even more in private than in public-that NASA served the purposes of national security; that NASA and DOD should act in the future, as they had in the past, on the basis of coordination, not joint management; and that the objective of military control in space required- the kind of research that was "being done at this stage more efficiently in NASA than it could be done elsewhere." When, in July 1963, Vice President Johnson phoned Webb to ask how much of "our present peaceful space program can be militarily useful," he replied,
The Gemini agreement raised troubling questions about the purpose and future of manned military space programs. Although Dyna-Soar was not dropped until December 1963, its cancelation had been expected for almost a year prior to the formal announcement. As early as March 1963, in discussions between top  officials of NASA and DOD, McNamara, according to Webb, had asked what benefits Dyna-Soar would have for NASA "and whether or not those benefits could not be obtained from some other program . . . rather than having to spend the 600 million dollars involved in funding the Dyna-Soar project."67 NcNamara was already thinking ahead to a military space station, and the summer of 1963 was marked by interagency negotiations on the coordination of studies for manned orbital space stations that led to a joint agreement in September. The announcement of Dyna-Soar's cancelation was yoked with one to authorize the MOL, a program that would use a Gemini capsule to house a two-man crew for a stay of up to thirty days. This led to the impression that MOL was a substitute for Dyna-Soar, when the truth was a little more complex. Dyna-Soar had attempted too many things at once, among which, to develop precise reentry and landing capabilities and to test the capability of man in space. It had outgrown the Titan I and Titan II boosters and, in any case, could carry only one man, rather than the two-man crew of MOL. But the latter was not simply a tradeoff for the former; Dyna-Soar was aimed at maneuverability and reentry problems, while MOL was to investigate the uses of men working in space in a "shirt-sleeves" environment.
The final authorization of MOL was piecemeal. The initial decision of 10 December 1963 was followed by the joint NASA-DOD agreement of 25 January 1965 and the final approval given by President Johnson on 25 August 1965. While many details remain classified, enough is known about MOL to clarify the reasons that led to approval. The elements leading to the 1963 MOL decision included pressure from the Air Force to demonstrate the need for a manned military space program and from Congress (especially the House Government Operations Committee) to explore the military uses of space, and the sentiment, not stated publicly by Defense officials, that the Air Force should receive something in return for the cancelation or phasedown of the Skybolt missile, the B-70, and Dyna-Soar.68 The significance of Dyna-Soar in this context is that there was not room in the Air Force budget for both Dyna-Soar and MOL, and that cancelation of the former was one more illustration of McNamara's concern for economy and cost-effectiveness. Because MOL would be based on proven technology, rather than on Apollo hardware not yet flown, the Air Force's case was much improved. By basing the program on Gemini rather than Apollo, the Air Force could also use its Titan III, rather than NASA's Saturn IB.
The 1963 decision to authorize MOL inaugurated a definition phase that lasted twenty months, to August 1965. The original plan called for the "Gemini B" capsule, attached to an orbiting pressurized cylinder (the "laboratory") that McNamara described as "approximately the size of a house trailer."69 Although the program would be supported by Gemini's tracking network, it would be totally funded and managed by DOD. Yet on certain critical matters, McNamara either said nothing or spoke in the vaguest terms. As DOD officials saw it, MOL was not a space station and did not come within the terms of the September agreement with NASA. Yet the reasons given for the program only raised doubts  as to its necessity. How accurate was McNamara's assertion that MOL involved no duplication of Gemini? What, precisely, could MOL do that could not be done equally well by unmanned spacecraft or by Gemini itself? And if MOL was not a space station (though to all appearances it seemed to be one), how would it affect NASA's plans for a post-Apollo space station? These doubts were not so much removed as ignored by McNamara and top Air Force officials as they began the tedious process of specifying the details of MOL.
In fact, this preliminary phase lasted as long as it did because the MOL concept raised questions of cost, timing, and, most of all, requirements. An ad hoc subcommittee of the Space Vehicle and Space Sciences Panels of PSAC reported in November 1963 that a general purpose space station could not be justified for military or nonmilitary purposes; that a nonrecoverable space station should not be considered by the military until it had been decided that the Gemini and Apollo capsules were inadequate for military purposes; finally, that DOD should begin a test program using a one-man Gemini capsule launched by a Titan II.70 Moreover, the Bureau of the Budget, which had to authorize the program, took more than a year to study MOL before approving it. As late as November 1964 the Bureau's Military Division could find little to justify a separate MOL. A Bureau staff paper of that date assembled most of the objections to MOL and summarized them forcefully:
And the study concluded that the MOL "did not now appear justified on the basis of the originally stated need for an experimental testing of the potential capabilities of manned space flight for high priority military purposes."72 If the Defense Department thought MOL necessary, then the program should, at the least, be "nationalized," that is, made the basis for experiments of both military and general interest. In short, elements within the Bureau of the Budget were still very sceptical about MOL twelve months after it had been initially approved.
What was the NASA position toward MOL? Webb, Dryden, and Seamans were told in advance of the MOL decision and agreed to it. Other NASA officials, however, were not briefed, not even George Mueller who, as Associate Administrator for Manned Space Flight, would be the official most concerned with the project.73 It was Webb's stated policy not to "second-guess" the military and not to oppose DOD programs that did not directly oppose NASA interests, if MOL could be said to fall into that category. It would also be politic, at the least, not to clash head-on with the Air Force over something that the Air Force was bent on having. Publicly, NASA officials defended MOL up to the day of its cancelation  in June 1969. They insisted that MOL was a military project within the overall national space program; that it complemented, rather than duplicated, Apollo; and that the two agencies were working on the Air Force project in the same spirit in which they had defined Gemini. 74 And indeed, during 1964 and 1965 NASA and DOD concluded a number of working-level agreements defining the MOL mission, several of which involved the good offices of the Director of the Bureau of the Budget.75 Privately, however, several NASA engineers and managers took a position scarcely distinguishable from the 1963 report of the PSAC Space Vehicle and Space Sciences subcommittee. As they saw it, the Air Force was maneuvering DOD "into a position of defining the MOL program as the National Space Station," thereby putting pressure on NASA to fly its experiments aboard Gemini B. Furthermore, the Air Force wanted to remove the program from AACB coordination and start a direct relationship with NASA.76 On technical grounds, they were equally sceptical; considerably less than 1 percent of the data obtained by MOL would be superior to what would be obtainable from available systems.77 On the basis of studies conducted by NASA for DOD between January and May 1965, the use of Apollo-Saturn was found preferable to Gemini B/ Titan III. The former could place very large payloads in orbit, which could be manned continuously through rendezvous resupply operations; while the Apollo command module could also accommodate a three-man crew, rather than the two-man crew of Gemini B. 78 In effect, some NASA technical managers suspected that MOL really did duplicate NASA programs, at a time when influential Congressmen were demanding less duplication and more standardization.
President Johnson approved MOL, one may assume, because the Air Force made a convincing case on its behalf, while those within the Government who opposed it fell into line. Sometime early in December 1964, the Director of the Bureau of the Budget (Kermit Gordon) and the President's Science Advisor (Dr. Donald Hornig) decided to support it. On 10 December Gordon recorded the understanding reached between himself, Hornig, Webb, and McNamara. It was agreed that the program would emphasize military experimentation, assembling structures in space, and reconnaissance; other objectives, mainly scientific, would be sought in cooperation with NASA; and NASA and DOD would define experiments to support these goals.79 The studies carried out by NASA as part of the agreement were intended to refute the contention that MOL would simply duplicate Gemini and Apollo. The pressure to cancel or drastically alter MOL had been growing since the fall of 1964; thus Senator Clinton Anderson (D-New Mex.), chairman of the Senate Committee on Aeronautical and Space Sciences, wrote to the President urging that military and civilian space stations be combined to avoid waste.80 The joint statements of Webb and McNamara on 25 January 1965 dismissed that possibility out of hand. Both confirmed the understanding arrived at in December: MOL would be a military program with NASA support; it would be directed to military purposes but also to certain broader scientific purposes; and NASA would compare Apollo and Gemini systems to see which would best serve the ends of the program.81 By April DOD had decided to proceed  with Gemini B and to let a design definition contract with McDonnell Aircraft. On 4 June the Military Operations Subcommittee of the House Government Operations Committee released a report that pushed the MOL in the most forceful way.82 The committee recommended that DOD "should, without further delay, commence full-scale development of a manned orbital laboratory . . . project." On 9 July NASA and DOD officals briefed the Space Council on MOL and NASA plans for manned space stations. Finally, on 25 August President Johnson approved MOL. After almost two years of discussion and analysis, the Air Force had its program.83
MOL was approved because the Air Force, backed by several congressional committees, wanted it, because NASA chose to support it, and because the Budget Bureau and the President's Science Advisor ultimately accepted the merits of the experimental program.84 The views of the decision makers on the merits of the classified experiments and on whether MOL could effectively implement them actually controlled the decision. If McNamara, Hornig, and the others had not been convinced, the decision would have gone the other way.85 NASA took the position that MOL did not duplicate any current NASA program, was not a space station, and could be justified on its own terms. Whatever reservations NASA officials had, they kept to themselves. Given the limited experimental nature of the final approved program, it could be argued that MOL did not compete with NASA's plans for manned orbital space stations. NASA's plans were far more grandiose. In any case, neither the Manned Orbiting Research Laboratory nor the National Multipurpose Space Station nor "Apollo X" was ever approved in anything resembling the original form. The threat to NASA was that the duplication argument could be turned around so easily. Prior to MOL approval, the program was criticized for duplicating NASA hardware and systems. Once approved, the argument turned back on NASA. Even before final approval, one NASA official wondered if pressure
And in March 1966 the House Government Operations Committee issued a report recommending a merger of MOL and Apollo Applications.87 Whatever the merits of the duplication controversy, the existence of a manned military program in space was bound to impinge on NASA's post-Apollo planning.
The connections between military and civilian space programs were so ramifying and complex that it is advisable to state the most important ones as summer: conclusions.
 First, although it began as an agency that depended heavily on DOD support, NASA succeeded in freeing itself from overt DOD control by 1963. Whether one considers Gemini, the management of the Merritt Island Launch Area, or the existence of colocated tracking stations, the pattern is the same. NASA would cooperate with DOD but never to the point of relinquishing its authority to meet s needs. The shift from dependence began within the first two years following NASA's creation. NASA began to work on active, as well as passive, communications satellites, despite an informal agreement with DOD to work only on the former. NASA had moved well beyond such understandings in asserting its right to manage its programs, to modify military launch vehicles to serve as boosters, and to let contracts to firms already heavily involved in defense work.
During NASA's first three years, the Air Force went to considerable lengths become the dominant partner in the national space program. Even some years later, Boone observed that "the Air Force is inclined to look upon NASA as a competitor rather than a partner in the field of space.'' 88 By 1963, however, the Air Force needed NASA almost as much as NASA needed the Air Force. NASA as doing basic research in the life sciences, in the composition of the upper atmosphere, and in propulsion that was as valuable for military as for civilian purposes. The NASA centers had research and test facilities that the services needed badly, among which were sixteen different wind tunnels, a ground-based glut-motion simulator, and an 18.3-meter vacuum environmental sphere. The framework within which the two agencies coexisted could accommodate almost any kind of program management, whether it was a program managed by one ;agency with the other sharing in planning experiments (Gemini, MOL); a joint program (X-15); a program started by one agency and transferred to the other (Syncom, the large solid motor); a joint program primarily funded by one agency (XB-70); or programs "whose success is dependent on the functioning of separate, -operating systems" (Syncom, Anna IB geodetic satellite). The relationship was strong enough to endure almost any strain, provided it was grounded in mutual specs.
Second, the inevitable overlap of two agencies working in the same research areas was made tolerable by the concept of warranted duplication. There could be duplication in programs, facilities, and research tasks as long as they were addressed to agency needs or represented different approaches to the same research problem. Two facilities or programs were seldom identical, in any case. But duplication could even be defended as a positive good, insofar as it widened the number of options available. The ballistic missile programs of the 1950s were, in that sense, test cases for the virtues of duplication. As Sapolsky observes,
 Third, most of the coordination between NASA and DOD was handled quite efficiently by working-level groups at the centers or by AACB panels. There were, however, cases where agreement could only be reached by the agency heads or, failing that, by referral to the Bureau of the Budget or the President. In one instance, NASA and DOD could not agree on funding arrangements for NASA launches at the Eastern Test Range. The history of that controversy suggests the old saw that the case was "hopeless, but not serious." When one considers that the matter was exhaustively studied and that the Director of the Budget could provide only an interim solution, it suggests that NASA and DOD could survive the strain. Even the most explosive disagreements, such as those over Gemini and the space station programs, were defused, if not by general managers like Seamans or R&D directors like Brown, then by the heads of the agencies involved. The important point is that the preconditions for coordination existed. What were they? One, that DOD should accept NASA's definition of a "coordinated" program as one where concurrence was "not required as a pre-condition to further action" 90; two, that both agencies should have a mutual interest in cooperating; three, that NASA's programs, particularly in manned spaceflight, should have a priority sufficiently high to require support by the services; and four, that both agencies should centralize the organization of their space and launch vehicle programs to make cooperation possible. The last point refers to the NASA reorganizations of 1961 and 1963, the creation of the Air Force Systems Command, the merger of the Eastern and Western Test Ranges under a single National Range Division, and the use by both agencies of AACB panels as standing committees to regulate the working-level coordination of military and civilian programs. Perhaps the most remarkable thing about the NASA-DOD relationship is not that it worked so well, but that so often practice was better than theory, and mutual interest overrode the funding and duplication controversies recounted in this chapter.
* The first stage of the Saturn rocket was developed by the von Braun team at the Army Redstone Arsenal at Huntsville, Alabama.
** Originally called the Atlantic and Pacific Missile Ranges, they were redesignated the Eastern and Western Test Ranges in May 1964.
*** In July 1967 the two divisions were superseded by the Headquarters Space and Missile Systems Organization (SAMSO).
**** Manned Space Flight, unmanned Spacecraft, Launch Vehicles, Space Flight Ground Environment, Supporting Space Research and 'Technology, and Aeronautics. The Civilian-Military Liaison Committee was formally abolished by the President's Reorganization Plan No. 4, 27 July 1965.
***** When Seamans replaced Dryden as cochairman of AACB.
****** However, Skylabs II, III, and IV, and Apollo-Soyuz, all of which used the Saturn IB, were from KSC.
+ True for the 6.6-meter motor but not for smaller boosters. Minuteman was a solid-fuel missile, while the Titan III-C used 3.05-meter solid-fuel strap-on boosters.
++ Under pressure from CAO, NASA and DOD carried out certain "economy studies" in 1968-1969, which aimed at consolidating such things as photographic work at the Eastern Test Range and the Kennedy Space Center.
+++ This was the justification for the Ames Space Science Research Laboratory, although some similar work was performed at the Air Force Cambridge Research Laboratory.
++++ E.g., the Spacecraft Instrumentation Evaluation Facility at MSC, whose capabilities could not be duplicated by existing DOD facilities.
+++++ As DOD considered the V/STOL Research Laboratory at Langley to be.
++++++ According to one source, McNamara proposed that DOD take over all manned flight in Earth orbit, NASA all flights beyond Earth orbit.
+++++++ This, however, was what NASA did for DOD in establishing a limited warfare program.