In terms of its proclaimed objectives, the civilian space program in the 1960s was an extraordinary success. The manned lunar landing was achieved on schedule, the record of successful launches improved steadily between 1961 and 1970, and NASA worked steadily and well to develop launch vehicles capable of being produced serially and of accepting a variety of payloads. As far as this book is concerned, the problem is to explain how decisions about NASA's organization accounted for this success-if, indeed, such success occurred because of the organization, not in spite of it. No doubt it is possible to state the reasons for NASA's success in general, straightforward terms. The argument might run along these lines: Congress gave NASA a blank check because the lunar landing had to be managed on a crash basis. Furthermore, by committing itself to extremely precise, measurable, all-or-nothing goals, NASA had powerful incentives not to fail.
All this may be true as far as it goes, but it does not go far enough. Terms such as "crash program" or "decentralization" derive their meaning from a specific context. For example, the 1963 organizational changes discussed in chapter 3 "decentralized" NASA so that lines of authority reverted to what they had been prior to 1961. But one cannot step in the same stream twice; the 1963 decentralization occurred in the context of the vast new lunar landing mission ordained for NASA by President Kennedy in May 1961. These conditions did not recur. Similarly, one should be cautious in applying the term "crash program" to Apollo. If the term is used to denote a program in which "the cost is not the major consideration . . . several parallel approaches are taken for the solution of each major problem, and . . . overtime and multiple shifts are utilized to a very large degree," then Apollo was not a crash program, since the first two criteria, at least, did not apply.1 Whenever these terms appear in this book, the reader should be aware of their contextual implications.
 The salient features of the space program must be sought elsewhere. Among them was the often fruitful tension between applied research and basic research. In the development of a space capability, space applications, supporting technology, and aeronautics, NASA's program was essentially applied research. The manned lunar landing was comparable to the development of a polio vaccine, rather than to a cure for cancer. The latter entails fundamental knowledge about the nature of the cell, whereas the former involves only a more sophisticated use of existing research techniques. On the other hand, space science was accorded a prominent position in NASA's program and organization. Three NASA centers were devoted-two of them almost completely and one of them substantially-to space science.
Other features of the agency's operations may help to explain the success of the space program. There was the ability of the agency to draw on the scientific community for advisors, principal investigators, and consultants; the absence of distinction between the budgetary process and short and long-range planning; the refusal of NASA officials to enforce a separation between those who built flight hardware and the scientists who designed and developed the experiments to be flown; the ability of the agency to draw on the capabilities of the Department of Defense for launch support, contract administration, program managers, and astronauts; and, finally, the authority vested in the Administrator to waive some civil service requirements in recruiting executives from the outside. But the decisions that made these features possible are not, in a sense, the most important administrative decisions. To use Nelson's distinction, these are decisions within an organizational structure rather than decisions about that organizational structure, which Nelson sees as the major decisions.2 Decisions about an organization occur when "a given organizational regime is limited in the range of contingencies it can handle effectively.... when circumstances evolve outside of this range, the symptom is a growing restiveness (on the part of one group or another) with the routine flow of events and decisions; and . . . successful resolution requires some kind of significant reorganization." It is this kind of administrative decision about the NASA organization that is discussed here.
Anyone who studies the morphology of large organizations is apt to be struck by the power of a few key administrative decisions to set the organization's course. By "administrative" is meant the choices between alternatives that shape the agency from within, not those imposed from without. The peculiarities of such decisions are that they represent choices among alternatives, some of which might have been equally successful; they tend to be cumulative in effect; one such decision requires many others, both to carry it out and to make explicit what was only implied in the original directive; they are at once shaped by and shapers of the agency's terms of reference; and they are not themselves discrete, self-contained decisions as much as they are bundles of administrative acts that gradually rise to the level of agency policy.3
Consider, for example, the NASA policy, avowed from the beginning, to contract out research and development rather than do such work entirely in-house.
 The decision itself was made before the mechanism needed to make it work-a rational source selection procedure had been perfected. One might go as far as to say that NASA began with an operating philosophy whose implications took some seven years to work out. In this regard, the decision to contract out stands somewhat apart from other decisions considered here. The others were formally enunciated in a specific directive as internal agency policy, and each contributed greatly to NASA's ability to interpret its mission. In this respect, administrative decisions differ fundamentally from those that establish the mission. The decision to prescribe phased project planning as NASA policy is different in kind from the decision to establish a civilian space agency or President Kennedy's decision to "send a man to the moon and bring him back before the end of the decade." The latter decisions are "organic"; the former, attempts to translate them into day-today operating procedure. As with most translations, something is inevitably lost in the process. First, management and line officials will carry out broad goals as they understand them. Second, the same communications gap as that between the agency and the executive branch is repeated within the agency itself between the offices that constitute it and the senior officials who act as links between the organization and an environment that may be indifferent or even hostile. No formulation or directive can cover all contingencies; no decision will be seen in precisely the same way by those affected by it; no policy is so impregnable that alternatives are inconceivable.
The decisions that follow were broad formulations; they had to be sold to the center directors, headquarters line officials, and the functional staff. In this respect, technical, political, and management considerations could not be separated for long. The decision to develop the F-1 and J-2 engines for the Saturn rocket stages could not be made in advance of the mechanism for procuring such hardware. Similarly, neither Apollo nor Gemini would have been feasible as agency programs in the absence of a firm policy that established relations between the centers and headquarters.
The decision to rely on private industry, rather than in-house staff, for development of NASA programs has probably been the key internal decision in the history of NASA, yet its genesis is by no means clear. The policy of contracting out was arrived at in gradual stages, each of which made the policy more precise, hence more effective. The Space Act authorized NASA "to enter into and perform such contracts . . ., and on such terms as it may deem appropriate . . . with any person, firm, association, corporation, or educational institution."4 Except for a proviso concerning small-business set-asides, the act (by its silence) gave NASA the freedom to work out its contracting procedures, the forms of which were borrowed from the armed services procurement regulations, adapted to the needs of a young, rapidly expanding agency. The key steps were as follows:
These steps made NASA procurement philosophy more precise and flexible, while placing heavy responsibilities on the centers and the line organization to initiate feasibility studies, prepare project guidelines, assist in the evaluation of contractor proposals, and supervise and evaluate work accomplished under contract. Despite its undeniable success, the policy of contracting out-begun by T. Keith Glennan, the first NASA Administrator, and greatly expanded by James E. Webb-raised serious problems that were at once technical and political.
First, there was the question of how large an in-house staff was needed to monitor prime contractors. If the staff was too small, lacked experience in managing large development contracts, or was inclined to favor contractors who had worked for the agency before, NASA would be at the mercy of its contractors. Should the staff be too large, there would be pressure from the centers to do the work in-house. There was always some tension between forcing the centers to learn to deal with contractors and setting aside a certain percentage of funds for small internal projects, without which NASA could not retain its most talented scientists and engineers. As Wernher von Braun, Director of the Marshall Space Flight Center (MSFC), once remarked to a House subcommittee:
Second, NASA officials had to decide what functions might properly be delegated. The negotiated contract might be for hardware, services, or study reports; the line between what was and was not proper was often exceedingly fine. For example, was it proper for a company to do a feasibility study, recommend that a project be authorized, and then receive a contract to produce hardware for the project? Should a company that offered technical advice to NASA be able to benefit from that advice? What kinds of decisions could not be delegated to private bodies of any kind?
A third problem had to do with contracting for specific services when civil service staff was available to do the job. This was the same problem in another form: What criteria should the Government use in procuring something, be it a launch vehicle, management of a center cafeteria, or test and checkout services? No set of criteria (the Bureau of the Budget issued several sets of guidelines) could meet every contingency.6 The presumption in favor of a commercial source, save when no such source was available or when using it might somehow disrupt the agency's work, smacked too much of a short-term solution. Excessive dependence on private industry would eventually endanger NASA's ability to determine the  kinds of programs it wanted, to estimate reasonable prices, and to evaluate program results. But too much work done in-house would have revived those problems it was the ostensible purpose of the contractual instrument to resolve: a lack of flexibility, the accumulation of a large permanent staff for whom jobs had to be found, and a concomitant vested interest in particular programs. No decision in NASA's history has been fraught with greater consequences than this one. No aspect of NASA organization was left untouched by it.
Another significant decision concerns the reorganization in November 1961, by which responsibility for hardware and responsibility for programs were concentrated in the same program offices. The essence of NASA program planning was organization by goal rather than by hardware. Prior to 1961 a separate Office of Launch Vehicle Programs was responsible for engine and launch vehicle development. If NASA was to be a single agency, rather than a congeries of semi-independent fiefdoms, hardware and programs had to be combined in the headquarters program office and the centers then reporting to them. The catalyst for the change was the decision to give highest priority to the manned lunar landing. What had been discussed for over a year within NASA now became a matter of urgency. The basic changes-omitting the details for now-were (1) that the centers were to report directly to the Associate Administrator, (2) that four new program offices were established in place of the four that had been abolished, and (3) that a new Office of Tracking and Data Acquisition was set up to provide agency-wide support.
The reorganization gave the Manned Space Flight (MSF) program the organizational prominence needed to carry out its goal and, in so doing, paved the way for the centers to manage the large development contracts on which the success of the program depended. From these two things-the special freedom of the Office of Manned Space Flight (OMSF) and the doubling and tripling of appropriations and manpower-resulted the most important changes of the next two years: the introduction of incentive contracts and the conversion of older contracts from cost-plus-fixed fee; the establishment of an Office of Industry Affairs with agency-wide responsibility for procurement; and the decision to build a research center dedicated to electronics, the crucial discipline in the refinement of space hardware. The 1961 reorganization, in sum, emphasized the "commonality" of agency objectives "The manned lunar landing was a NASA objective, not just the objective of the Office of Manned Space Flight. All NASA field installations were to contribute to its accomplishment, not just the centers labelled as manned space flight centers Agency-wide functions . . . were to be performed for the benefit of the entire NASA program, not just one segment of it."7
A third decision is that of the November 1963 reorganization, after which the field centers once more reported to the program offices, not to the Associate Administrator. This decision signaled the maturity of the NASA organization. The centers now had the capacity to conceive projects, develop specifications for contractors, supervise contractors, and conduct sufficient in-house work to maintain the excellence of their staffs. The priority of the MSF program was clearly recognized and with it the necessity of shifting the locus of decision making down  ward from the Associate Administrator to the heads of program offices, center directors, and project managers. Many factors had to converge before such a system could work or be worked. First, there had to be standardized procedures for authorizing projects, tracking resources through the agency, and translating data received into information about ongoing programs. Second, management needed techniques to measure performance, costs, and schedules. Third, NASA needed special scientific and engineering personnel for its research and development programs. It is perhaps truer to call the 1963 reorganization an end rather than a beginning, because without the capabilities listed no such dispersion of authority would have succeeded. In other words, the functions of senior management, i.e., Administrator James E. Webb, Deputy Administrator Hugh L. Dryden, and Associate Administrator Robert C. Seamans, Jr., would be less day-to-day supervision and more long-range planning, handling interagency matters, and attending to the delicate international implications of the U.S. space program. The 1963 reorganization did not change NASA's organizational philosophy; it ratified it.
The fourth major decision was to go to "all-up" flight testing of Saturn vehicles (1963-1966). The all-up philosophy had a rather long history before rising to the level of agency policy. All-up testing means that a vehicle "is as complete as practicable for each flight, so that a maximum amount of test information is obtained with a minimum number of flights." 8 In other words, all three stages of the Saturn V were to be flown with the Apollo command and service modules instead of being tested and flown separately. Unlike the decisions listed so far, it may seem more technical than administrative; again, unlike the others, it was made by George E. Mueller, Associate Administrator (OMSF), and the OMSF Management Council, rather than in the Administrator's office.9 If the MSF program, by the size of its appropriations and the extent of its resources, was the pacing element of NASA planning, if Apollo was the pacing item for that program, and if the lunar landing had to be completed before the end of the decade, then any decision that would decrease the number of test flights would be crucial. In one respect the all-up decision was like the previous decisions discussed: It evolved from earlier decisions and, in turn, presupposed subsequent decisions to implement it. The all-up decision presupposed the July 1962 decision to use lunar orbit rendezvous as the mission mode for Apollo, and it required an ever stricter control of quality and monitoring of contractors, the budgeting of weight within the launch vehicle and spacecraft (itself requiring major advances in electronics and miniaturization), much closer attention to improving some portions of the Apollo program without delaying other portions, and the building in of redundancy wherever possible. All-up testing falls into the category of administrative decisions because it was developed by agency managers and center directors who felt that small incremental development steps would be too time consuming and expensive, and because it did as much as any single decision to establish the allocation of resources and manpower within NASA.
Another significant decision was to strengthen the Office of the Administrator (December 1965-March 1967). Here, too, the decision to create an Office  of the Administrator was inextricably involved with prior top management decisions The year 1965 had been one of intense activity, and the capacity of Webb and Seamans to absorb information, to see the agency as a whole, and to make sense of the welter of reports and issuances generated at the lower levels was stretched very thin. Following Dryden's death on 2 December, Seamans, while remaining general manager, became Deputy Administrator. Webb then moved to tighten his control over the functional management offices. In a 29 December memorandum for agency-wide distribution, he announced the establishment of an Office of the Administrator as "a single, uncompartmented entity to afford maximum, direct, personal and informal contact between Dr. Seamans and myself and our associates."10 In addition to detailing the functions of the secretariat, Webb's memorandum was the most complete expression of his philosophy of the administrative division of labor up to that time. Seamans remained in charge of day-to-day operations; Webb, of protecting NASA from an ever more threatening external environment. The functional staff serving the new office would monitor and assist the program offices.
The second phase of Webb's strategy, begun in early 1967, changed drastically after the Apollo 204 fire of 27 January. The Office of the Administrator had made the division of functions more clear-cut and precise; by itself, this could go only so far to reduce bureaucratic inertia: the delays in reaching decisions, the shifting of decisions upward, the chaotic state of the management issuance system, the real possibility of an Apollo "stretch-out," and so on. Webb appointed Harold B. Finger, manager of the NASA-Atomic Energy Commission Nuclear Propulsion Office (and leader of a task force to study NASA organization) as Associate Administrator for Organization and Management. His functions would include supervising the functional offices, such as Industry Affairs and Administration, and acting as the central point of control within the agency for all project authorizations. Instead of the relatively loose control exercised since 1963, Webb, through Seamans and Finger, worked to bring about much tighter control from above. To many, the Apollo fire was understood to be the failure of North American Aviation, the prime contractor for the Apollo command and service modules, to do what it was supposed to have done. To Webb, it meant more: He had been misinformed by his staff to the point where such an accident had become first possible, then likely, then seemingly inevitable. Finger's mission would be to make sure that the program offices, particularly OMSF, got what they needed and that Webb and Seamans would know it. In effect, the Office of Organization and Management became for NASA what the Bureau of the Budget was for the executive branch.
The decision to establish the project approval document as the single basic device for authorizing and directing NASA programs (1962-1968) marked Webb s final move in controlling agency resources. Phased project planning could break a project into four distinct but related steps, allowing management the option of continuing or terminating at three precise points. What was still needed was a document that could match program costs against agency resources, present and future. The budget, after all, was not just an expression of agency desires; it  had to show, line item by line item, the relation between programs, reporting systems, and changes in midcourse. To this end, the system projected by Webb would include an operating budget and the project approval document (PAD), which would contain "the official statement of objectives, technical and management requirements . . . allocated resources . . . and other controlled milestones.... An absolute one-for-one ... relationship would exist at all times between controlled line items . . . in the NASA Operating Budget and the related PAD's, whereby a change in either will require a concomitant change in the other."11 Although PADs had been the principal mechanism for authorizing the use of appropriated funds since the early 1960s, they were now to be the single device for authorization, with only one or two signatures required for approval.
To a degree, any list of key decisions must partake of the arbitrary. A decision to do one thing is also a decision not to do something else. Indeed, some of the most important NASA decisions for the period under consideration were those not to do something: not to ask for the $400 million supplemental appropriation requested by OMSF in 1963, a request that might well have unbalanced the NASA program; not to set up a general advisory committee, as a high-level scientific panel recommended in 1966; not to change the prime contractors for the Apollo spacecraft in the aftermath of the January 1967 fire.
Several more technical decisions could be listed. But at the risk of tautology, the ones listed here are the key decisions because they mattered greatly to the future of the space program. They were not imposed from without, nor were they the result of attempts to placate this or that group through a shift in organizational window dressing. The nature of the decisions does not presuppose one specific organizational theory; the policy analyst may find in any or all of them elements of "disjointed incrementalism," group consensus, participative management, or the theory of organization as problem-solving coalition. The value of the list is that it invites the reader to consider themes, certain of which are latent in the facts presented and others that are reserved for fuller treatment in subsequent chapters. The themes are, first, the sequence in which the decisions were made, for example, the way in which program authorization procedures were developed before, just after, and subsequent to the lunar landing decision; second, the way that political, technical, managerial, and administrative considerations came to form a body of decisions; third, the distinction (if any) between gross structural changes in the organization and the many smaller changes whose net effect may be as great; and finally, the contrast between the success of NASA programs and the frequent inefficiency of the agency's administrative housekeeping. The discussion now turns to an account of NASA's founding and the problems and achievements of its first eleven years.