Managing the Unmanageable

The United States showed the world how a government could foster the freedom required for basic science research, the results of which can be negligible or earthshaking. That freedom—and America’s technological lead—have eroded at the same time


IN THE FALL OF 1942 J. ROBERT OPpenheimer walked into the Presidio, in San Francisco, to undergo a physical. He was about to be tapped by General Leslie Groves, the head of a top-secret Army project, to establish and administer a classified laboratory in Los Alamos, New Mexico, that would investigate the possibility of developing an atomic bomb. Oppenheimer and his bosses took it for granted, given the source of the money, the lab’s directive, and the need for security, that the facility would follow conventional practice and be militarized. Hence his trip to the Presidio; the physical would be the first step toward being commissioned a lieutenant colonel, which in turn would be the first step toward taking command of the largest and most ambitious scientific project yet attempted.

But Oppenheimer never went through officer training. He never needed to. Key scientists refused to come to Los Alamos under conditions of military hierarchy and bureaucracy, which, they said, were antithetical to the spirit of science. Scientists thrived when they were judged according to competence, not rank. Under pressure, Groves relented. On February 25, 1943, he sent Oppenheimer a letter promising that although the military would provide resources and general direction, for at least the time being Los Alamos would remain civilian, managed by contract with the University of California. Even Groves would come to recognize that this arrangement worked far more effectively than the original, conventional one would have.

The letter set a momentous precedent for relations between the U.S. government and its scientists. When, after the war, the newly founded Atomic Energy Commission set up a string of national laboratories, it chose to manage them according to a similar scheme, called administrative contracting, in which independent managers were hired, often universities or similar nonprofit organizations. The system was viewed as essential to the special environment needed for basic research, which, unlike applied research, seeks an understanding of the structures of nature for its own sake. Partly as a result, the United States forged ahead of other nations in postwar science and in the technology that derived from it.

Today that innovative and unique system is in jeopardy, threatened by a tendency to manage federal undertakings of all kinds, including scientific projects, as though they were businesses—constrained by business procedures and requirements and responsive to business incentives and pressures. Although we will focus on the problems of the national laboratories, owing to our familiarity with them, we believe that similar management problems are or soon will be experienced in connection with large government-sponsored scientific projects in other branches of science. The impact on science has gone virtually unnoticed by the public, because it has occurred in a series of tiny steps, no single one of which has been large enough to attract attention. But it is oppressively evident to those who work inside the national laboratories. The attempt to treat basic research as a business has slowly changed the structure and spirit of these laboratories, and begun to strip away the special protection this country once accorded them and to taint the fragile atmosphere needed for them to thrive. At a time of widespread lamentations about the loss of U.S. technological competitiveness, it is ironic that we are destroying one of the most important means by which we established that technological competitiveness in the first place.

Changing the World by Chance

ALTHOUGH BASIC AND APPLIED RESEARCH ARE OFten intertwined in practice, their intellectual missions are distinct. Basic research aims to recognize previously unknown structures of nature, whereas applied research aims to make some already known process possible or more effective. The structures of nature recognized in basic research may have practical applications, but whether they do or not, and if so of what type, are generally not the professional concern of the basic researcher. At the beginning of this century a popular postprandial toast around the Cavendish Laboratory of Cambridge University, in England, was, “The electron: may it never be of any use to anybody!” That wish, as the physicist Abraham Pais once wryly remarked, was unfulfilled. A few decades earlier, during an inspection of the Royal Institution, in London, a skeptical William Gladstone questioned the potential value of electricity. “Sir,” came the waggish riposte, “someday you will tax it!”

The American sociologist Robert Merton once devoted an essay to what he called “the unanticipated consequences of purposive social action.” Human action, he wrote, often brings about processes or results that are unintended by the actors, and these can affect either the actors or society as a whole, and can be beneficial or not. Although Merton was specifically discussing social action, his point is equally true of scientific research. But whereas most human action seeks to avoid unanticipated consequences, basic research courts them. Many discoveries are unsurprising, expected outcomes of deliberate research programs. Nevertheless, basic researchers are always aware of the possibility of—and often hope for— novel developments. The British theoretical physicist P.A.M. Dirac, who devised the “Dirac equation,” which accounts in a comprehensive way for the behavior of the electron, used to say that his equation was smarter than he was, because it contained solutions to problems of which he was unaware. Moreover, historical illustrations of unintended practical consequences of discoveries are legion. We will cite a few examples to illustrate the mundanity of unforeseen consequences in discoveries of genuinely new structures of nature.

Improving medical techniques was hardly on the mind of the German physicist Wilhelm Roentgen as he tinkered with cathode-ray tubes in November of 1895. But when he explored the curious fact that a fluorescent screen near his apparatus was glowing, contrary to all expectations, he ended up discovering a wholly new phenomenon of nature, which he called x-rays. Within three months they had been used to examine bone fractures. His discovery triggered a series of other important scientific discoveries, such as that of radioactivity.

“I don’t think the idea of helping suffering humanity ever entered our minds,” Howard Florey, of Oxford’s Dunn School of Pathology, once recalled of the moment in the late 1930s when he and his colleague Ernest Chain began a survey of antibacterial mechanisms. But on their list of microbes to study was penicillin, which had been discovered accidentally only a few years before but whose antibacterial properties had been neglected. Much of modern medicine is based on a substance whose discovery and development were matters of sheer chance and disinterested academic research.

“Moonshine,” responded the British scientist Ernst Rutherford, who discovered the nucleus in 1911, to the suggestion that energy might be obtained from it. Atomic energy was equally far from the thoughts of the Italian physicist Enrico Fermi when, in 1934, he began bombarding the nuclei of all known elements with neutrons, and was puzzled by the results he achieved with the heaviest known element, uranium. The German scientists Otto Hahn and Fritz Strassmann were trying to make sense of Fermi’s odd results when, at the end of 1938, they announced that barium was a by-product of such bombardments. A month later two other German researchers, Otto Frisch and Lise Meitner, were only trying to explain that puzzling result when they suggested that uranium nuclei could split, or “fission,” with a concomitant release of energy. The outcome of all this puzzle-solving, it hardly needs mentioning, transformed the world.

The unintended consequences of scientific research can reach far beyond science and technology, to have a cultural impact as well. One remarkable and well-documented example is the profound effect that telescopic astronomy had on Milton’s poetry, and in particular on his imaginative depiction of the space of the universe in Paradise Lost. Some three hundred years later one might consider how x-rays, CAT scans, and radio telescopes have further altered human perception and imagination. It is unfortunate that the beneficial effects of scientific discoveries often become so thoroughly integrated into the world that they are taken for granted, as part of its warp and woof, whereas the pernicious applications of scientific discoveries are often portrayed as representative of scientific activity itself. Pessimists anticipate that basic research will deliver new forms of power to be socially abused; optimists anticipate a reform of social structures which will eliminate such abuses.

Temperamentally, scientists tend to be optimists, and if pressed will defend their work by saying that in the long run, at least, the effects of science are beneficial. But arguments about the value of basic research that are based on anticipations of either social utility or harm are weak, because they are based on analogy: the future will be like the past. A deeper motive for basic research is that it ultimately leads not just to understanding of the structure of nature but to self-understanding as well. What is nature? and Who are we? are not distinguishable questions. “Whatever nature has in store for mankind, unpleasant as it may be,” Fermi once said, “men must accept, for ignorance is never better than knowledge.”

An Odd System That Worked

THE PHRASE “MANAGING BASIC SCIENCE” MIGHT at first seem oxymoronic, like the oft-repeated jest about “military intelligence.” Management is the effective coordination of resources and personnel toward a particular end; how can one coordinate an activity whose end is unforeseen and unforeseeable? Nevertheless, basic science stagnates when it is not effectively supervised. Salaries of basic scientists, for instance, are lower than those of applied scientists. Basic scientists cannot be paid competitive salaries, because there is no guarantee that their work is marketable and even when it eventually happens to be so, the returns are usually too far in the future to affect them. Nevertheless, many extremely bright people find that the relative freedom to follow one’s nose in preparing a research program, collaborating and competing with smart and dedicated colleagues, and the thrill of the prospect of making fundamental discoveries are ample compensation for unequal pay.

Creating and maintaining a healthy scientific culture is the aim of science management, and it involves tending to both the intellectual and the institutional conditions of basic research. The intellectual condition is freedom of inquiry—allowing and even encouraging scientists to follow new paths should they open up, to risk dead ends, to shift direction abruptly on a hunch. The institutional condition of basic science is a laboratory environment that facilitates such freedom of inquiry. A laboratory is more than a collection of equipment and the space to house it. It is a theater in which experiments are performed and witnessed, and like theaters of the more familiar kind, it is specially built for that purpose. Providing such environments was less complicated in the first few decades of this century, when scientists worked alone or in groups of twos and threes, when important laboratory skills were glassblowing and carpentry, and when one required no more than an ordinary room in which to perform an experiment. At that time laboratory equipment was workbench-sized and relatively inexpensive, and much of it was built by the researchers themselves; until the 1930s the first morning task for a physicist who needed power supplies often was to build batteries. With the Manhattan Project and the postwar establishment of the national laboratory system, however, all that changed. “Little Science" became “Big Science,”in the words of Alvin Weinberg, a former director of Oak Ridge National Laboratory, in Tennessee. Researchers might work in teams not of twos or threes but of dozens and even hundreds; the cost of experiments reached not tens of thousands but hundreds of millions of dollars. Making equipment was subcontracted to industry, and the space to house it might require hangar-sized buildings.

Today Weinberg’s phrase is unfortunately associated with an unproductive debate about science funding policy in which “Big Science” often carries connotations similar to those of “Big Business.” The suggestion is that large scientific projects unfairly monopolize scientific capital, squeezing out the little guy who might make valuable innovations if given a chance. But the analogy is false. Knowledge generated by large scientific projects, unlike the profits of large corporations, becomes the property of the entire community and restructures the scientific background against which research teams large and small plan and execute new ventures. Moreover, “Big Science" is a general term whose meaning varies from one branch of science to another. It can refer to the construction of large instruments used by only a small fraction of the community at once, as in astronomy; large equipment complexes serving many individuals simultaneously, as in materials science; or the coordination of the work of numerous small research teams, as in biology. In each case the funding needs, the size and role of research groups, and the information flow between project users and the wider scientific community is different. It is thus meaningless to debate the value of Big Science in general; projects must be judged on an individual basis.

Nevertheless, the management problems of such largescale projects are similar, and therefore the experience of high-energy physics, in which such problems were first felt, is likely to be emblematic. Big Science meant that high-energy physics outgrew the environment single universities could provide. The national laboratories had a special mission: to provide favorable environments in which such science could continue to grow. But the new scientific theaters had vastly different requirements from those of just a decade or so earlier. It was not simply a matter of bigger equipment, along with the need for more-extensive construction and planning. The new laboratories also required things like offices for procurement, maintenance, health and safety, security, architectural planning, and budget. They needed departments for photography, technical information, public information, and legal counsel. The commitment to maintaining such a large organization inevitably posed a threat to the freedom and flexibility of the basic research that was supposed to be carried out. Science management had become a problem—one that grew with the size of the laboratories.

The solution adopted was the administrative contract. Administrative contracts date from the era of the Atomic Energy Act of 1946, and the historic compromise it effected between the U.S. government and the U.S. physics community. The government was trying to balance its desire for a first-class scientific program with its desire for secrecy and control of the direction of nuclear-reactor technology. The physics community wanted the nuclear reactors and particle accelerators that were soon to become fundamental tools of Big Science, but also wanted to work independent of the government, which was the only conceivable source of funds. Administrative contracts permitted the government to participate in science while preserving the atmosphere of university laboratories.

Administrative contracts were soon written establishing several national laboratories, some through the reorganization of existing labs: Brookhaven National Laboratory, on Long Island, New York (the contractor was Associated Universities, Inc., a nonprofit group of universities in the region); Berkeley Radiation Laboratory, in Berkeley, California (University of California), now called Lawrence Berkeley Laboratory; Argonne National Laboratory, outside Chicago (University of Chicago); and Oak Ridge National Laboratory (built largely by Du Pont and now run by Martin Marietta). In addition to these basic-research facilities, a string of laboratories oriented toward military projects were also set up under administrative contract. Besides Los Alamos (University of California), these included Livermore Laboratory, in Livermore, California (also the University of California), now called Lawrence Livermore National Laboratory; and Sandia Laboratories, in Albuquerque, New Mexico (AT&T, through a subsidiary, the Sandia Corporation).

When the contractor was a university or similar nonprofit organization, it received a management fee to cover its costs. Commercial companies came to receive an “award fee” along with reimbursement for their costs, in an amount depending on a judgment about the effectiveness of the management in any given year. The commercial companies participated not for the money but partly in order to contribute to the good of the country, partly to provide a training ground for employees, and partly to benefit from the transfer of new technology. The Sandia Corporation, for instance, has managed the Sandia Laboratories for forty years without any award fee.

The arrangement worked so well that when a group of European nations founded CERN, a laboratory in Geneva, in 1953, its management was patterned after the administrative contracts of the Atomic Energy Commission (AEC). “It is the desire of the Commission,” a typical contract stated at the outset, “to procure for the Government managerial skill and responsibility which will permit flexibility in administrative controls and freedom from detailed supervision.” Many of these contracts carried what became known as the sweetness-and-light clause: “It is the intent of the Commission and the Contractor that this agreement shall be carried on in a spirit of partnership and friendly cooperation with a maximum of effort and common sense in achieving their common objectives.”

The special character of administrative contracts can best be seen by comparing them with the two most common types of government contracts, fixed-price and costplus-fixed-fee. Suppose, for instance, that the government wants to buy Army hats. It draw’s up specifications, determines which sizes and colors it wants, and then solicits bids. If the winner of the contract, a fixed-price one, cannot produce the specified hats on budget, it must bear the additional cost itself or renegotiate the contract. Cost-plus-fixed-fee contracts cover situations where the scope of the work precludes a basis for determining a fixed price. Suppose the government wants a series of metals to be evaluated for use in space vehicles. It might produce a detailed description of the tests it wants done, prescribing how long the process is to take, how often it will want to receive reports, and so on, and invite interested parties to negotiate a contract in which the government agrees to pay whatever costs are incurred plus a fixed fee. In both kinds of contracts the aim is to pin the contractor down to as many specifics as possible. Both presuppose a conventional buyer-seller relationship.

Administrative contracts envision an entirely different relationship between contractee and contractor, one that is essentially collaborative. It is taken for granted, for instance, that the relationship will be long-term; contractors are formally reviewed after five years but do not necessarily have to undergo a competitive rebidding process. Although the AEC determined its laboratories’ basic direction, approved long-range plans, and played an important role in health and safety policies, other management issues were the responsibility of the contractor, who was allowed maximum flexibility in deciding them.

The Intrusion of Bureaucracy and Politics

ADMINISTRATIVE CONTRACTS WERE CREATED IN the knowledge that special management methods are required by the special environment in which basic research thrives. Yet the fruits of basic research—and their obvious ability to transform society—have made it tempting for the government to think that basic research can and should be managed for social ends, and that therefore conventional management models and methods can and should be applied to it. Little by little the government has succumbed to this temptation, and has come to adopt the perspective that basic-research facilities can be operated just like other corporate entities. The result has been to erode two fundamental kinds of independence originally granted to contractors: the relative independence from bureaucratic restrictions applied to other federal agencies, and the relative independence from political pressures.

The federal government long ago established a system of regulations to cover the contracting and purchasing functions of its agencies and any subcontracting carried out by those agencies. The reasons for doing so include the government’s interest in preventing collusion, fraud, incompetence, and inefficiency, and its interest in promoting certain social ends.

The first administrative contracts, however, exempted the laboratories from such provisions. The promotion rules, salary structures, and personnel regulations of civil service were considered inappropriate to the laboratory environment. Because scientists on the research frontier need to be able to respond quickly to new developments, the elaborate accounting and procurement practices required of government agencies were also thought inappropriate. For such reasons, the contractors who managed the labs were bound by few of the restrictions that were standard in other government contracts, except for basic health and safety provisions. That is not to say that the contractors were given carte blanche; they were accountable for their work through periodic reviews whose conclusions were made available in publicly released reports. But the contractors were freed of the ordinary bureaucratic burdens placed on federal agencies.

That freedom did not last long. Within a few years of the signing of the first administrative contracts many of the boiler-plate provisions of standard government contracts were being written into administrative contracts. These were not unreasonable, and contractors did not find them burdensome. But severer restrictions were applied, through a process known in the jargon as “flowdown,” by which restrictive requirements and regulations in one contract tend to flow into others. The first flow-downs came through the AEC, which wrote procurement regulations into its standard contracts. It was often unclear both to the laboratories and to their Washington contract supervisors which of the AEC regulations had to be followed by the laboratories in subcontracting. The problem was exacerbated by the fact that the ultimate court of appeal for contract disputes with the government is the General Accounting Office, which had the right to examine all AEC contracts and subcontracts to evaluate their legality.

Over time the GAO’s authority grew, as did the number of its regulations and the penalties for breaking its rules. Inevitably contractors began to play it safe and adopt the conventional contracting practices even when not technically necessary.

AEC procurement regulations were but one source of flow-down. Around 1960 a second source appeared with the creation of a set of federal procurement regulations. Once again, it was frequently unclear both to the laboratory managers and to contract administrators which regulations were to be applied to the administrative contracts; once again, the fact that the GAO had the final say discouraged the laboratories from making aggressive and flexible contracting decisions.

In 1975 the AEC ceased to be, and most of its functions were taken over by the short-lived Energy Research and Development Administration, which in turn was superseded by the Department of Energy, in 1977. The flow-down now became a torrent, thanks to a change in the character of the agency overseeing the laboratories. The AEC staff had always included a number of scientists who engaged in basic research, and the agency’s five commissioners had always included at least one basic scientist. The professional staff of the AEC, like those of the Tennessee Valley Authority, the FBI, and a few other agencies, was exempt from civil-service requirements, meaning that it had a great deal of flexibility in selecting and assigning laboratory personnel. The Department of Energy, however, has never had a basic scientist at its helm, and is staffed largely by officials with little or no familiarity with basic research and its particular requirements. People arrived at the new agency with no awareness of the reason why administrative contracts had been established in the first place and no experience in handling them, and expected that the contractors who ran the national labs would follow the same regulations as everyone else.

In 1982 an advisory panel reported to the DOE that “the laboratories have become grossly overburdened with detailed reporting and other paperwork requirements, the utility of which frequently is not apparent and which unnecessarily divert resources from their research and development missions”; nothing, however, was done in response to the complaint. Similar conclusions were reached in subsequent years, with similar outcomes, by a presidential commission and a second advisory panel.

The result is that the national laboratories have to divert an increasing portion of their resources in order to satisfy federal bureaucratic provisions from which they were meant to be exempt. It is a question not of science administration having to grow in scale with the size of scientific projects but of the breakdown of the independence that the administrative-contract system was designed to safeguard. Three decades ago the procurement regulations that bound laboratories were available in a small booklet. Today each lab has several feet of “DOE orders,” with additional ones arriving weekly. In the past ten years alone the budget departments at the national laboratories have doubled—not because they need additional help in preparing the budgets but because of the DOE’s neverending requests for information. Whereas agreements between contractors and the government aimed at a relationship of “sweetness and light” forty years ago, today the relationship is described as “at arm’s length.” Some examples:

• The DOE requires laboratory directors to approve each request for international-travel funds, and if the requests are filed within thirty days of the start of the proposed trip, they must be approved by the appropriate DOE offices in Washington. This regulation is absurd. Given the international character of contemporary science, international conferences are necessarily one of its important tools. Brookhaven scientists, for instance, make about 300 foreign-travel applications a year. DOE regulations require so much reporting and follow-up that the Brookhaven administration includes two people who work fulltime to process applications.

• Following a recommendation of the 1982 advisory panel, the DOE established a program to provide lab directors with seed money for promoting new initiatives. The program has led to a number of important new research efforts that otherwise would have been delayed or abandoned. Last year, however, some scientists at Lawrence Livermore National Laboratory attempted to use the program to develop an idea to visit Mars using inflatable spacecraft. The plan completely bypassed input from NASA. Angered, Congress retaliated by eliminating the new-initiatives program. Ultimately it was restored, but with additional procedures and restrictions. Whatever one’s opinion of space exploration, the story illustrates a bureaucratic tendency to punish all labs for an action at one.

• Last year Brookhaven couldn’t get its integrated circuits made. Custom semiconductor chips are a component of experimental devices and instrumentation of all sorts. For a laboratory to have no new chips is something akin to a hospital’s running out of penicillin. Integrated circuits are prohibitively expensive for a laboratory unless bought jointly with other institutions. When Brookhaven joined a consortium of others to acquire chips, however, some officials read the DOE requirements as prohibiting such a cooperative venture because of certain restrictions on submitting a purchase order to private industry. The laboratory spent a year tinkering with the agreement to make it valid, only then determining that the officials were merely playing it safe with the DOE requirements. Before the matter was straightened out, several major projects ground to a halt.

Revitalizing Scientific Culture

IT WAS ONE INTENTION OF THE ATOMIC ENERGY ACT to keep political hands off the basic-research budget as much as possible. One part of the act, for instance, provided for the evaluation and selection of research proposals by independent panels of scientists. Another sought to insulate the selected proposals and expedite the budgetary approval needed from Congress by the creation of a single congressional committee to oversee the basic-research budget, the Joint Committee on Atomic Energy—the only permanent joint committee with continuing legislative responsibility ever created. Many in and out of Congress were jealous of the joint committee’s tremendous power. But throughout its tenure the joint committee respected the independence of the laboratories in conveying to the scientific community merely general areas of government concern, and insisting on the peer review of programs meant to address these concerns.

The demise of the AEC meant the eventual dissolution of the joint committee, which left the basic-research budget to be parceled out by a complex network of committees with overlapping jurisdictions. The absence of a single committee with a continuing vision exposed the basic-research budget to the political whims of successive presidential administrations. Laboratory programs established by one administration have been axed by the next.

Moreover, the parceling out of the basic-research budget also meant a growing vulnerability to pork-barreling, which can be defined as any non-peer-reviewed and -approved project. During the lifespan of the joint committee a few minor pork projects had made it into the budget, but the committee had aggressively kept the lid on. The lid opened, however, during the Reagan Administration. In 1982 George Keyworth, the presidential science adviser, apparently at the request of certain California constituents, attempted to bypass the peer-review process and insert funds for a National Center for Advanced Materials (NCAM, but soon referred to as NSCAM) directly into the DOE budget as a “presidential initiative.” A $140 million project, it would have been the largest undertaking ever in materials science. A storm of protest led Congress to defer the project temporarily, but much of it has been reinstated.

After that episode Congress lost the restraint with which it had traditionally approached the basic-research budget. If presidential initiatives were possible, it was argued, so were congressional initiatives, and universities began to lobby Congress directly for them. In 1983 Cassidy & Associates, a lobbying firm, succeeded in getting Congress to earmark $5 million of the DOE budget for a chemistry building for one of its clients, Columbia University. The same year the firm snagged another $5 million for a vitreous-state research laboratory at Catholic University, in Washington, D.C. Cassidy has become the lobbyist of choice for universities. It is estimated that in the four-year period from 1982 to 1986 the funds earmarked for pork projects bypassing the peer-review process soared from $2 million to $236 million.

Behind the proliferation of bureaucracy and politics is a deep misunderstanding of the scientific process, in which basic research is viewed as essentially a corporate undertaking and hence something that can be manipulated for profit or social ends. Basic research flourishes— and society reaps the greatest benefits—when it is viewed not as a profit-making venture or as an instrument of social change but as an exploration of nature. To provide basic research with the special conditions appropriate to its execution amounts not to granting it a special status but simply to recognizing what it is and what are the conditions under which it should be done.

The hidden cost of imposing on basic research the same procurement procedures, budgetary constraints, and general regulations imposed on other areas of the federal bureaucracy is the sharp reduction of that flexibility which keeps basic research vital. We appear to be losing our technological leadership, and if we are not careful, we may lose our scientific leadership as well. It is true that the United States still holds the lion’s share of science Nobel Prizes awarded since the Second World War. But we may be riding on past accomplishments. The 1990 Nobel in physics was won by three Americans for work done two decades ago, the 1989 prize by an American for work done in the 1940s, the 1988 prize by three Americans for work done a quarter century ago. In the past decade most of the physics Nobels awarded for recent efforts have gone to Europeans.

Recently the Department of Energy has taken some promising steps. In the current year’s budget, for instance, it has recognized basic research in science as a category in its own right, which it calls “Fortifying Foundations”; in previous years the budget for basic research in science was classified under “Energy Research & Development.” But much more needs to be done. We propose the following steps:

• Return to the original administrative-contract idea of flexibility and independence. For instance, the number of congressional committees through which the basic-research budget has to pass should be limited, and many accounting and procurement requirements eliminated. One cannot run science as a procurement activity.

• Establish a long-term science policy for basic research, identifying areas of interest and opportunity, and couple this with a two-year funding plan. The long-term policy would prevent the basic-research budget from fluctuating and make programs more effective by buffering them against the whims of Congress and of new administrations. A mechanism for establishing such a policy already exists, in the form of national science advisory and DOE committees; their mandate should be extended. Two years of funding cannot be guaranteed, since Congress makes only one-year appropriations; nevertheless, planning funding according to that time frame would be an improvement over the present system. Since congressional project additions are a way of life, they will continue. However, they should be budgetary add-ons and not come at the expense of ongoing programs.

• Reduce the number of DOE orders to laboratories, and end the bureaucratic mentality that formulates a new rule following every offense. Much more effective would be to treat the individual case first. Each DOE institution— and, indeed, each of the national laboratories—is unique. Each has different facilities, skills, styles, personnel, problems, and goals. A solution to a problem at one lab may not work and may even be harmful at another.

Genuine scientific culture, like all human culture, is achieved through organic growth rather than the execution of a plan. One cannot suddenly decide to buy a forefront scientific program and then go out and exploit it. Do we demand of basic research that it be of potential social or military use? Do we insist that it pay for itself? Or do we try to foster in those who do science the ability to follow their intuitions about how the world works? On the answers we give to such questions depends the future of science in the United States.