The Satanic Verses celebrates hybridity, impurity, intermingling, the transformation that comes of new and unexpected combinations of human beings, cultures, ideas, politics, movies, songs. It rejoices in mongrelization and fears the absolutism of the Pure. Mélange, hotchpotch, a bit of this and a bit of that is how newness enters the world.
Salman RushdieFootnote 1The transnational turn has created the space for a rich variety of new and stimulating questions for historians. It is, in fact, astonishing to see what happens when we decentre the nation state as the dominant unit of historical analysis.Footnote 2 Comparative national studies subvert uniqueness and exceptionalism. Non-state actors like NGOs and foundations, and international organizations like the United Nations and the World Health Organization, populate the historian's landscape with new entities that force us to interrogate the limits of the power of the nation state and to position it vis-à-vis other instruments of social change. Freed from the subterranean chains that bind them to national containers, historians begin to see ‘ideas, things, peoples and practices which have crossed national boundaries’. Their language reaches for ‘metaphors of fluidity’, and they talk of circulation and flows through networks that bind together a transnational community of actors.Footnote 3 All that is solid melts into air.
Transnational history poses specific intellectual challenges when the United States is at the heart of the analysis.Footnote 4 On the one hand, through much of its history America has been enmeshed in global interrelations. On the other, for most of the twentieth century at least, the United States has been a major power that could aspire to mould the world in its own image, rather than being obliged to accommodate itself to the world in which it found itself. Driven by a sense of historical mission to enhance America's global reach – and thanks to the accumulation of multiple instruments of power at this one pole – the United States appears to be ‘the clearest embodiment of the idea of the self-sufficient nation-state’, as Akira Iriye puts it.Footnote 5 Transnational history seeks to upend that idea by inserting America in the world. It sees it as one node in an interconnected global network of circulating ideas, people and practices – but a pole that rises above the others, and that sometimes dominates them.
The renewed interest in networks, travel and circulation in the history of science is symptomatic of this determination to break free from the national frame. Indeed, James Secord has recently suggested that ‘how and why’ knowledge circulates is ‘an issue of real analytical significance – in fact, the central question of our field’.Footnote 6 To answer that question for the period that is of interest here – the Cold War – we have to bear two crucial considerations in mind for they bear directly on how knowledge flows are structured. First, as already suggested, the United States was the dominant scientific and technological power on the globe until at least the 1970s. Second, the politicization of science and technology, and their enrolment in the agenda of the national security state, imposed constraints on the circulation of knowledge. The pursuit of American pre-eminence required that a balance be struck between encouraging the transnational flow of sensitive knowledge – interconnectivity – and a retreat behind high security walls – autarky.Footnote 7 The post-war, networked, knowledge world was not flat; it was lumpy, its topology defined by the enormous concentration of scientific and technological power at one pole and the United States’ determination to regulate knowledge circulation to sustain its competitive advantage.Footnote 8 It is often remarked that the transnational approach, in seeking to transcend the national framework, and to stress interconnectivity, movement and flow, risks obliterating the nation state altogether. This obliteration is all the more inappropriate when the United States is one of the social actors. As Marilyn Young has noted, ‘De-centering America in one's head is a good thing. But it does not of itself create a world free of its overwhelming economic and military power’ – nor of its immense scientific and technological weight after 1945, one might add.Footnote 9
An exploration of the circulation of knowledge in a lumpy world can learn much from the vibrant new work on science and empire or, more generally, on the relationship between ‘centre’ and ‘periphery’. This work creates the space for a history of science that recaptures the dynamic interaction between diverse and interconnected sites of knowledge-making.Footnote 10 By framing the analysis in terms of transnational networks rather than by the boundaries of the nation state it challenges a model that sees knowledge flowing from an initiating ‘centre’ in one dominant country or region to a receiving ‘periphery’ in another. It disrupts the associated unidirectional logic implied by the production → diffusion/circulation → imposition/appropriation sequence (even if the latter grants agency to the ‘receiving end’). This logic maps power and knowledge as vectors arching outward from a national centre driven by a hegemonic, not to say imperialist, urge. By contrast, a transnational approach invokes a schema that stresses interconnectivity and interdependence, derived from mutual interaction and feedback.Footnote 11 Production and circulation are woven together into a seamless web. For Kapil Raj, instead of a linear, unidirectional model of cross-border knowledge flows, the encounter between ‘centre’ and ‘periphery’ – and even these categories have been subverted since it often only makes sense to talk of multiple centres – is now seen as a face-to-face transaction between agents from different poles in which they collaborate in the co-production of scientific knowledge in response to specific needs and interests.Footnote 12
This paper builds on these insights to explore both the ‘how’ and the ‘why’ of knowledge circulation in the face-to-face encounters between British and American nuclear scientists and engineers who were developing gas centrifuges for uranium enrichment in the 1960s. These questions have different, if intersecting, answers and engage different, if intersecting, social actors and bodies of literature. The ‘how’ directs us to the material practices of the scientists themselves, the mundane procedures that constituted the coconstruction of what I call hybrid knowledge through exchange or circulation in laboratory settings. The ‘why’ situates the laboratory in the world. It foregrounds the motives that the stakeholders had for cooperating in the first place. In this paper it embeds the encounter in the trajectory of the Anglo-American ‘special relationship’ in civil and military nuclear affairs that gained momentum from 1958 onwards, a trajectory that was constituted by reciprocity. As John Baylis has put it, once the British government had persuaded the American administration to remove the restrictions on nuclear exchange imposed by the McMahon Act in 1946, the core of the ‘special relationship’ between the two countries ‘was maintained primarily as a result of reciprocity rather than sentimental attachment or vague notions of kinship’. Co-production was only possible, and was only sustained, as long as both partners were persuaded that they could derive some useful new knowledge from it.Footnote 13
Reciprocity is driven by the determination to enhance the national knowledge base; that motive can also disrupt the encounter to the point of terminating it. To see this we need to ‘avoid the amorphous quality of an anthropology of flows and fragments’ and deploy a ‘network concept [that] puts as much emphasis on nodes and blockages as on movement’.Footnote 14 The much-needed destruction of a rigid centre/periphery distinction has sometimes come at the cost of recognizing that knowledge often flows in an asymmetric, regulated field of force constituted by differences in power between its various poles. This was certainly the case here. Anglo-American cooperation in nuclear matters may have been constituted by ‘reciprocity’ from the late 1950s on, but it was a reciprocity sought desperately by the British and modulated by the Americans, who could determine the terms of an encounter almost at will. As Andrew Pierre put it, from the late 1950s Britain often ‘tied herself to the American order of priorities in research, development and production and in some respects at least became partially dependent on American satellite intelligence, navigation and radio communication systems’.Footnote 15 The construction of Britain's ‘independent’ nuclear deterrent owed much to combining indigenous initiative with American resources, and required such a dovetailing of British policies with American priorities that London could easily be held hostage to Washington's demands.Footnote 16 It is important to realize that the encounters between scientists and engineers from both sides of the Atlantic that will be described below generated hybrid knowledge through reciprocal exchange in an asymmetric field of force in which the United States had the upper hand thanks to its vast nuclear complex put in place after the Second World War.
The ‘how’ and the ‘why’ of knowledge circulation are captured in the notion of hybridity. Hybridity is one of the key concepts in transnational history. It is usually used to describe transformations in the cultural identities of migrant peoples who crisscross the borders of nation states. Rushdie's characterization captures the messiness and complexity of the process of knowledge-in-the-making that my sources reveal, pointing to those multiple modes of interaction that constitute the material practice of the encounter. But, given the context in which he uses it, it also hints at the mingling of national identities that is central to the concept of hybridity in a transnational context.Footnote 17 All knowledge is admittedly ‘hybrid’, made up of bits and pieces of this and that, insofar as it is the product of a social process. But here another dimension is added: the national provenance of those discrete elements in the transnationally produced melange. By understanding the knowledge embodied in the gas centrifuge as hybrid in this more specific sense, we gain purchase on the determined attempts by scientists and engineers on both sides of the Atlantic to define ‘British’ or ‘American’ elements in the hotchpotch of knowledge that they produced in the late 1960s. And we more readily see that the affirmation of the national in the hybrid is also the performative expression of power.
The context of knowledge co-production
In the early 1960s scientists and engineers working for the United Kingdom Atomic Energy Authority (UKAEA) became intrigued by the possibilities of enriching uranium using gas centrifuges spinning at very high rotational velocities.Footnote 18 By 1967 they were convinced that this technique would be about 25 per cent cheaper than the gas diffusion process that had been originally developed in the Manhattan Project. The government, advised by the UKAEA, promptly cancelled plans to reactivate the gas diffusion plant at Capenhurst, which had been temporarily mothballed, to produce fuel for its second generation of nuclear power reactors.Footnote 19 Britain would use instead the revolutionary new technology for the extension of the facility near Chester.
The UKAEA were not alone in their enthusiasm for the centrifuge. In spring 1968 Dutch and German industrialists announced that they, too, were developing this revolutionary technology.Footnote 20 Wary at first, the UKAEA soon perceived this as an opportunity. They were planning to build a pilot plant to test a prototype centrifuge that had emerged from eight years of research and development (Mark I). The United Kingdom could lever the lead that it had thanks to Mark I, along with its experience of managing an enrichment plant, to shape a tripartite collaborative project with the Continent. Whitehall saw an Anglo-Dutch-German venture as exemplifying the best features of the European technological community that Prime Minister Harold Wilson hoped to construct under British leadership. It was also an opportunity to enhance the United Kingdom's credentials as a partner worthy to enter the European Economic Community (EEC).Footnote 21
There was one major obstacle to this scheme. Between December 1960 and early 1965 teams from the UKAEA had collaborated on centrifuge research and development with colleagues at Oak Ridge working for the United States Atomic Energy Commission (USAEC). Their shared starting point had been a breakthrough in centrifuge design made by an Austrian émigré at the University of Virginia, Gernot Zippe. Zippe published his initial results in the open literature in 1960; all further research was immediately classified.Footnote 22 The collaborative effort between British and American scientists and engineers was accordingly regulated by Article IX (c) of the 1955 US/UK Agreement for Co-operation on the Civil Uses of Atomic Energy. This article required that ‘no material, equipment, device, or restricted data’, and ‘no equipment or device which would disclose any restricted data’, could be transferred to another entity ‘without the written consent’ of the party from whom it was received.Footnote 23 The USAEC was sympathetic to the British plans to collaborate with Continental partners. But they insisted that no ‘restricted data’ that the Americans had shared with them during their earlier collaboration could be passed on to the Dutch and the Germans without their written permission.
The UKAEA, with strong support from the chief scientific adviser, Sir Solly Zuckerman, argued that there was no breach of Article IX (c) since Mark I contained no restricted data of American provenance. The engineers and officials of the USAEC insisted that, on the contrary, it did. Meetings were held in 1969 in both London and Washington to resolve this dispute. During these debates engineers from the UKAEA and the USAEC opened the black box that was Mark I, and retrospectively reconstructed their mutual contributions to the device. The issue was so controversial that detailed minutes were kept to serve as an official record agreed upon by both parties. I will analyse these documents in several steps, beginning with an account of what transpired in the angry exchanges in 1969. This account will provide a rich platform for analysing the ‘how’ and the ‘why’ of the circulation of knowledge between scientists and engineers from the UKAEA and the USAEC in the 1960s.
Opening the black box, deconstructing the British centrifuge
A few technical details are needed before we proceed. The two most important parameters affecting the separation power of a centrifuge rotor are its length and the rotational velocity of the wall.Footnote 24 The output of enriched material increases directly with vessel length and, ideally, with the fourth power of the wall velocity (in practice, roughly with the square). The limits on wall velocity are set entirely by the stress properties of the materials used, which deform as the centrifugal forces increase. Long rotors, while desirable, have natural frequencies of vibration that are easily excited if the rotor is unbalanced at high speed. Balancing is never perfect, and if the vibration becomes violent enough the rotor and its bearings are destroyed. As it flies apart, the rotor can do extensive damage to other centrifuges connected to it in a cascade.Footnote 25 The economics of centrifuge development thus require extensive ‘life-testing’ to see how long a rotor can function before it disintegrates due to metal fatigue. One also has to choose between sacrificing output by operating subcritically (that is to say, at a wall velocity below a rotational speed that would excite the natural vibration frequency of the rotor), or operating supercritically. For supercritical operation one has to damp vibrational instabilities as the rotor accelerates through one or more resonance frequencies (for example, by introducing bellows into the wall of the vessel). All of these aspects of centrifuge development arose in the encounter between engineers from the USAEC (Paul Vanstrum and Myron Krantzer)Footnote 26 and from the UKAEA (Ned Franklin and Don Peirson) when they met face-to-face in London in March and in Washington in May 1969.Footnote 27 Their question was whether there was restricted data of American provenance in Mark I.
Rotor materials were the first bone of contention. The US team was studying the use of metal–fibreglass composites in the mid-1960s. By May 1963 they had such a rotor running in uranium hexafluoride (HEX) for a year without any ill effects. The UKAEA team, by contrast, had concentrated on using metals only (e.g. titanium alloys). By 1965, however, when the collaboration ceased, the British were reporting results in which they had used fibreglass. According to Vanstrum, this interest showed that American work had ‘contributed to the change of direction in the United Kingdom programme’.
The British emphatically refused this. Yes, the partners had discussed materials together in the early 1960s. Fibreglass was one of many options considered. Each had originally chosen different technical solutions – the United States working with composites, the United Kingdom with metal alloys only. The British changed direction, not because of anything in particular that they had learnt from the Americans, but simply because their titanium alloys were being seriously corroded by HEX. They selected aluminium wound with fibreglass as a ‘logical alternative’. The choice once made, the Capenhurst team had proceeded to make extensive ‘life-tests’ with fibreglass composites ‘which owed nothing to United States information’. Indeed the United States, far from helping the United Kingdom, had actually impeded matters. Franklin suggested that the Americans were just as interested as they were in the data on metals-only rotor bodies, and that a division of labour had been jointly agreed upon that had actually retarded rather than facilitated British progress (a charge that Vanstrum rejected outright, of course).
The overall thrust of the British argument was that they had learnt nothing specific about the use of materials from the United States. The solution they had adopted ‘was based on broad scientific and engineering considerations’. Anyone familiar with the problem would eventually decide to use composites. The use of composites was a ‘logical’ step once one had found that metals alone were seriously corroded. The addition of fibreglass was just one alternative (the use of carbon fibre was another). This the Americans could not accept. As AEC commissioner Gerald Tape put it, ‘the concept of using fibreglass was not unique: but the decision to select it in preference to other material and to base further development on it was another matter’, and inspired by the American work. The UKAEA argued that it had independently changed the direction of its research with no specific American input. The USAEC argued that the very fact that they had changed their research trajectory was thanks to United States work on composites.
Franklin also emphasized how much progress they had made after the United States withdrew from the collaboration in 1965. Mark I used different materials, its dimensions were different, its rotor had been significantly simplified, and the machine would be run subcritically. Life-tests were promising, with failure rates per annum below 10 per cent. A great deal of effort was also being devoted to production problems, including the best ways of connecting individual centrifuges into a cascade and of managing centrifuges that spun out of control using crash brakes. In sum, Franklin insisted that the new knowledge of centrifuge performance obtained by the UKAEA in developing a reliable production prototype owed nothing to the earlier research done with the United States.
Vanstrum objected. He pointed out that during the US–UK exchanges a wide range of centrifuge dimensions had been considered. He remarked that ‘much detailed data’ had been shared on the particular problems and geometry of the fibreglass windings. He stressed that ‘reliability was covered at some length during the exchanges’. He noted that American experience with the use of fibreglass-wrapped machines in cascades ‘had been reported to the United Kingdom in 1963’, that UKAEA visitors had been ‘taken on tours of United States laboratories containing cascades’ and that ‘data on cascade efficiency as well as United States cascade experience had been given to the United Kingdom’. Where Franklin stressed originality Vanstrum stressed the ‘interweaving of ideas’. In fact Vanstrum insisted that even if Mark I was ‘totally different’ (sic), ‘it would be influenced by the basic design data which had been exchanged’. Restricted data of American provenance would always be embedded in the ‘British’ centrifuge no matter how far it evolved beyond the early stages of research.
The UKAEA did admit that two features of their centrifuge that had been discussed with the Americans departed significantly from the original Zippe unclassified design.Footnote 28 One was the lower bearing. This had been simplified for economic reasons. The UK had adapted an American concept, and used a system of three pillars to support the spinning rotor, rather than the ball-and-magnet device preferred by Zippe. This had been patented, and was unclassified. The second, far more controversial, area was the so-called end cap.
The length of the body of a centrifuge contracts when it is spun at high speed (the Poisson effect). To compensate automatically for this deformation of the rotor, the UKAEA engineers fitted an end cap over each cylinder. This ‘lid’ was not flat but slightly conical or ‘dished’. The flattening of the cone at high rotational velocities offset the shortening of the cylinder, so keeping it aligned between the upper and lower bearings. End caps were not foreseen by Zippe. The UKAEA admitted that the idea of dishing the end cap fittings had been suggested to them by the Americans during their collaborative phase, but only in passing – ‘in fact during a fifteen minute conversation’.Footnote 29 Thereafter the British had worked out the details themselves, so they said. They had no access to secret American reports describing the theory. Instead they had independently drawn on knowledge that was well known in other fields of high-speed rotating machinery to compensate for the Poisson effect, and had applied it to the particular case of the gas centrifuge. Their design had also evolved as they adapted their prototype for mass production. For the UKAEA, then, the end cap, while novel, had simply been mentioned in passing by the USAEC as a way to improve centrifuge performance at high speed.Footnote 30 No ‘vital pixel’ (to quote Galison) of information on end caps had passed between the two partners that could be regarded as restricted data in terms of Article IX (c).Footnote 31
The American engineers refused this interpretation. Where the British spoke of a conversation of a quarter of an hour, they spoke of ‘a great deal of discussion’. While the British dismissed the significance of the exchange by claiming that no ‘unique’ design feature had been transmitted to them, the USAEC insisted that ‘it was difficult to conceive that some of this discussion was not reflected in the ultimate UKAEA design’. In fact Vanstrum and Kratzer remarked that, in the midst of the UK–US collaboration, they had taken out a classified patent in Britain on their end cap solution. How, then, could the British suggest that this could be readily devised by any competent engineer: the patent application was ‘evidence that the United States regarded their “end-cap” solution as unique and patentable as early as 1962 and that it was important to the centrifuge technology’. Of course the use of a conical cap to compensate for the Poisson effect at high rotational speeds was generally known: ‘it was the linking of this simple design solution to this particular problem which was important’, said Kratzer, echoing Tape's argument over the use of fibreglass. For the Americans the very idea of using an end cap on a gas centrifuge was novel, patentable and restricted data. It was disingenuous of the British to suggest that it could easily be derived from general principles, and wrong to imply that it was not a ‘unique’ contribution that the USAEC engineers had made to their thinking.
The exchange with the USAEC on 4 March persuaded the British that their attempt to dismiss as insignificant any knowledge acquired during the earlier collaboration had failed. The next day Zuckerman proposed to go further, and to provide the Americans with a ‘statement’ that ‘would specify, in general terms, the materials, the general dimensions of the centrifuge, and the general nature of its bearing and end-cap’. This document would also explain why the British did not believe that any information in their centrifuge could reasonably be regarded as ‘restricted data’ in terms of Article IX (c). Tape was not satisfied. He wanted Franklin to describe the Mark I prototype ‘in greater detail’. He wanted specifics. In fact he thought that it might be a good idea if the UKAEA arranged for United States personnel to ‘see the United Kingdom gas centrifuge project at first hand’.Footnote 32
This request for visual access caused immense consternation in the United Kingdom. They believed that the life-tested design of the end caps that they now used in their Mark I centrifuge gave them a technological advantage over both the US and their European partners. It was of considerable commercial and political significance, and was not to be given away easily. On 3 May 1969 Zuckerman went to Washington to present the British case against visual access to the USAEC.Footnote 33 USAEC chairman Glenn Seaborg, along with commissioners James Ramey and Francesco Costagliola, were present. Gerald Tape was also in the room, although he had officially left the commission just a few days before. The United States position had hardened considerably. Ramey quickly made it absolutely clear that the AEC had to have access to the British Mark I production model if they were to reassure the Congressional Joint Committee on Atomic Energy (JCAE) that no American information would be passed to the Dutch and the Germans. What did visual access to the Mark I prototype mean? Tape and others explained. It ‘would not involve taking the model apart screw by screw and item by item’. But it would mean ‘more than just a visual view’. ‘Experts who knew something about the technology [would have to] examine it’. This investigation would not extend beyond looking at an individual Mark I centrifuge. The British would not be asked to divulge details of the life-tests, nor would they have to explain how they engineered the device for mass production.
That was not all, though. Ramey insisted that there be a ‘firm procedure for informing the Americans in advance of what was to be transferred to the Dutch and Germans by way of advanced R and D for future models’.Footnote 34 This way the USAEC could check that no restricted data of United States provenance was being shared in the subsequent ‘Marks’. Tape put a time limit on the constraint: he surmised that, as research progressed, by 1980 it would be ‘hard to find a connection with the 1960–65 exchanges’. The British were being told to get permission from the USAEC before being allowed to pass on improvements to their Mark I prototype to their European collaborators. Every significant advance they made for the next decade (as it happens, until the time that the USAEC was planning to commercialize centrifuge enrichment itself) would be evaluated by their scientists and engineers and could, of course, be exploited to its advantage by the United States.
A Cabinet committee chaired by Zuckerman discussed the implications of this with John Hill, the chairman of the UKAEA, and one of his senior engineers, D.G. Avery, on 20 May.Footnote 35 They realized that visual access to Mark I did not mean a quick look. ‘The Americans would expect to be able to take our machine apart and to examine the pieces.’ Given that both teams began from Zippe's basic design, and that the idea of the end cap had been suggested by the researchers at Oak Ridge and discussed with them ‘many times’, a detailed examination of the British device would ‘inevitably’ reveal a ‘large number of points of similarity’. The USAEC ‘inspectors’ would be accompanied by lawyers who would stress these points of similarity (rather than the distinct contributions made independently by the UKAEA engineers), leading to ‘further argument and recrimination’. Hill concluded that, on balance, it was better to produce a reasoned argument against visual access than to open the prototype to inspection and give the Congressional JCAE ‘more information with which to cause trouble’.
There was a notable slippage in the UKAEA's position at this meeting. Until now they had claimed that they had only had one brief conversation about the end cap with the engineers from Oak Ridge. This single fifteen-minute exchange had now expanded to many discussions. Hill was asked to dig up the records. His findings were conveyed to Zuckerman the next day. Sir Solly had just penned a memo claiming that the British end cap was based on an independent application of a general engineering principle, that ‘no documents or “blue-prints” were exchanged on this specific issue’, and that ‘it would be immensely difficult to distinguish between specific US classified information and our application of basic principles of physics and engineering’.Footnote 36 Hill's new findings were a bombshell. The concept of the end cap had first been mentioned in 1962. In 1963 the US disclosed the dimensions and performance of an end cap four inches in diameter. In 1965 ‘two detailed papers were presented by the U.S. dealing with the stress analysis and lift of a dished end cap and reporting on their prototype’.Footnote 37 The main difference between what the British had learnt in these exchanges and what they did subsequently came down to developing a rotor with a diameter of four rather than five inches. The next day Denis Healey, the Secretary of State for Defence, demanded that there be an official enquiry.
The enquiry, which was led by Lord Penney, reported ten days later. It confirmed the worst. It noted that at an
especially significant meeting with U.S.A. representatives in May 1963, they showed the British team a drawing of their deforming end cap for a 4 in. diameter cylinder, together with experimental results on the deformation of this type of cap at various rotational speeds for various angles of the dishing cone and for various thicknesses of metal. They also told the British team of their theoretical work on the stresses and deformation of the end cap, and showed a drawing which included a dished baffle plate [dished diaphragms within the rotating cylinder that served as gas baffles].
The principle, the proportions, and the material of the dished end cap on Mark I were the same as those described to the British team by the Americans. Its geometry incorporated only ‘minor modifications’ of the American model.Footnote 38 The encounter had not been limited to a brief conversation. The UKAEA had learnt about the end cap in the course of a ‘fairly lengthy exposition, followed some two years later by the transmission of US papers’.Footnote 39 The exchange went well beyond basic research. It extended ‘from laboratory experimentation up to and including pilot plan design and operation’.Footnote 40
In a few short months the UKAEA's account of the encounter between researchers at Capenhurst and Oak Ridge regarding the design of the end cap had exploded from a fifteen-minute discussion to an extensive sharing of knowledge on centrifuge R & D from bench to prototype to production. At the meeting in London with the USAEC early in March, Myron Kratzer had insisted, against UKAEA protests, that in terms of Article IX (c) of the 1955 agreement ‘the United Kingdom could not include the conical end-cap in any exchanges with the Netherlands and Germany’ without explicit American permission.Footnote 41 Penney's three-man panel concluded unequivocally that the British design of the end cap incorporated US restricted data. Kratzer was vindicated.
On 3 June Sir Solly was back in Washington to meet with representatives of the USAEC (including commissioners Glenn Seaborg and Wilfred Johnson, as well as Myron Kratzer) and of the State Department. His hands were still tied by a UKAEA that was deeply reluctant to let American researchers into British labs. Zuckerman proposed to show the Americans an end cap and other items in Mark I that they thought might be derived from their research, but not to grant them visual access to the whole centrifuge. Regarding future developments, the UKAEA offered to give the USAEC an ‘outline’ of what they intended to share with their Continental partners, and to consult with them if and only if the British deemed it necessary.Footnote 42
Both proposals were summarily rejected. Nothing short of visual access to Mark I would satisfy the JCAE. Visual access would also be necessary in the future, if the USAEC requested it, to check whether any American restricted data was included in later production marks. The UKAEA capitulated. In an aide-memoire of 3 July 1969, Her Majesty's Government (HMG) agreed ‘to grant to the U.S. authorities visual access to their first production model’. As regards future R & D, HMG agreed to provide the US authorities with ‘an outline of the original U.K. contribution to the initial collaborative Research and Development programme’ with the Netherlands and Germany. They also agreed to consult in advance with the United States before transferring information to their partners. This consultation on future developments would include ‘access, if requested by the United States authorities, to the specific data or features’ which the Americans suspected might violate the terms of Article IX (c) of the 1955 agreement.Footnote 43
When Zuckerman met with the USAEC there was some discussion of the composition of the team that should come over to look at the Mark I centrifuge. To mollify the UKAEA, Seaborg and others suggested that it should be small – say, some three people. Although it would be quickest to send a team from Oak Ridge, ‘it might be possible to find people who were not concerned with current American centrifuge development at the working level’, or who had been engaged in exchanges in 1965 but were no longer active in the subject.Footnote 44 In the event the four-man team that visited Capenhurst and Risley for two days in mid-July had a good deal of in-house expertise among its members. It was led by Commissioner Theos Thompson, and it included Paul Vanstrum from the centrifuge team at Oak Ridge, one other centrifuge expert, and Nelson Sievering from the State Department.Footnote 45 The UKAEA scientists and engineers collaborated fully. Thompson told Zuckerman that his team was impressed by the amount of ‘independent development work done in the United Kingdom’. The discussions and presentations had dispelled ‘various technical doubts raised in the US’.Footnote 46 For Thompson the end cap dispute was now resolved, though of course he warned Sir Solly that the JCAE might deem otherwise. The dispute was officially closed some time later, on 1 October.Footnote 47 Six weeks later the British, Dutch and German governments put the final touches to a treaty establishing a new European uranium enrichment organization called Urenco.Footnote 48
The UKAEA now moved proactively to deal with the American demand to have access to their next generation of centrifuges. To avoid a repetition of the Mark I debacle the UKAEA proposed to put the onus on the USAEC to decide if any data they planned to share with colleagues across the Channel were restricted. To that end the ‘AEA proposed to show their whole field of second generation development work’ to the Americans. Early in March 1970 Commissioner Theos Thompson was back in Britain, where he visited Capenhurst and Aldermaston. He spent two days ‘to satisfy himself personally about those aspects of our centrifuge R and D programme which we shall be disclosing to our two partners’. Just before leaving he told Sir Solly that ‘as far as he [was] concerned, all [was] well’.Footnote 49 It seems that from that point on the British were no longer beholden to the USAEC. They moved ahead with their tripartite arrangements unimpeded by fears that the US would stop them in their tracks.
The ‘how’ of knowledge circulation/co-production
The probing enquiries by officials of the USAEC into just what the UKAEA scientists and engineers had learnt from researchers at Oak Ridge provides a rare and invaluable insight into the practice of knowledge-making in face-to-face encounters. As I pointed out earlier, our inherited conceptual apparatus for understanding knowledge flows is burdened by unilinear models that decouple production from circulation, that parse circulation in terms of transfer, and that treat reception as passive imposition or, more recently, as active appropriation. Raj's suggestion that we focus on face-to-face encounters and understand them as collaborative exercises in the co-construction of knowledge helps us to overcome these epistemological obstacles. His focus was the engagement between European ‘science’ and indigenous knowledges in South Asia, the contact was between different modes of understanding and different cultures, and the places of encounter were heterogeneous milieux, from trading ports to mountain passes, from drawing rooms to urban colleges. This case study shifts our attention back to the laboratory, the bench, the drawing office, the workshop, the filing cabinets and the conference rooms. It also deals with two communities that share not only language, but also skill sets, research practices, tacit knowledges and modes of civility (including an understanding of the political and legal constraints that surround classified collaboration). Their exchange is premised on reciprocity, on the assumption that each party has something to learn from the other, and that the trust that underpins collaboration can morph into the distrust that goes with competition.
The knowledge that the actors had of the centrifuge as they came together was subtended by a shared pool of publicly available information, most notably Zippe's design of the centrifuge, along with ‘broad scientific and engineering considerations’ or ‘basic principles of physics and engineering’ that any competent practitioner knew. Taking that as their starting point, the UKAEA's opening gambit was to interpret Article IX (c) as restricting only ‘specific reports received under the agreement, or drawings, or the transcribed content of notebooks’ between 1960 and 1965.Footnote 50 Thereafter, they said, the article could gain no purchase since their work had been directed to specifying, building and then multiplying a prototype, as opposed to simply trying to know more about the performance of centrifuges under a wide range of operating conditions. Their own detailed enquiries demolished the first argument. Gerald Tape would not hear of the second, arguing that only by 1980 would it be ‘hard to find a connection’ between the British production centrifuge and the exchanges that had occurred between 1960 and 1965.
These exchanges between British and American scientists and engineers were achieved by a dazzling variety of procedures, as we have seen. There were laboratory and plant tours. There were brief discussions and extended presentations. Detailed research reports, drawings and blueprints changed hands. There was physical ‘access’ that involved more than just an ‘outline’ or a ‘visual view’ or the possibility of carefully examining a component of Mark I, but less than reverse engineering by taking the prototype apart ‘screw by screw and item by item’.
It is noteworthy that, in effect, the UKAEA hoped to restrict the scope of knowledge covered by Article IX (c) to Latourian ‘immutable mobiles’ – research reports, drawings, extracts from lab notebooks and so on.Footnote 51 The USAEC refused any such restriction outright, as we have seen. Latour's model deals with the physical displacement of material supports; it describes how knowledge remains stable as it travels over space and in time. It is too thin to account for the practice of knowledge transfer in the face-to-face encounter between researchers from Oak Ridge, Capenhurst and Risley. For that we are better served by the ‘mundaneness postulate’.
Steven Shapin has drawn attention to the variety of ‘mundane features of social scenes’ that are involved in truth-making and truth-upholding. These include ‘conversing, persuading, cajoling, coercing, manipulating, testifying about experience and receiving the testimony of others, making marks on blackboards, inscribing in print and circulating printed objects’.Footnote 52 The ‘mundaneness postulate’ is a far more useful guide to the question ‘how did they do it?’ than are more general terms like transfer, reception and appropriation (and their vehicles, immutable mobiles). It cuts like a knife to the core of the material practices by which scientific knowledge is made. It also elevates the face-to-face over action-at-a-distance as a primary site for the analysis of circulation/production. Indeed it collapses circulation into production.
The ‘why’ of knowledge circulation/co-production
What was the point of these exchanges, the why of knowledge circulation? Most obviously, to learn what one could from the other. Reciprocity was a sine qua non for the possibility of the transatlantic collaboration happening at all. It enabled scientists to share design data and experimental findings and to converge on desirable technical specifications, to work more effectively by exploring different solutions to the same problem (for example, using metal alloys or fibreglass for rotor materials), to help one another ‘link’ a ‘simple design solution’ to a ‘particular problem’, and so on. It also presupposed a roughly equal level of expertise. The capacity to grasp what one learnt in an encounter and, above all, to capitalize on it in future depended on one's own state of knowledge. An engineer who knew something about centrifuges, but who was no longer working in the lab at Oak Ridge, would learn less from a visit to Capenhurst, Risley or Aldermaston than a colleague who was productively engaged in research; a lawyer would learn even less, and indeed would ask quite different questions. The breadth of participants’ knowledge determined how much they could learn from visual access alone and whether or not they needed to take the centrifuge apart piece by piece.Footnote 53 One has to be smart to be a spy.
The ‘why’ of knowledge circulation involves more than just the wish to learn, however. As my story shows, lurking beneath the active collaboration based on trust between 1960 and 1965 lay a competitive urge that burst into the open in 1969. The aggressive demands for access by the engineers and officials in the USAEC, and their dogged determination to force the British to reveal the latest details of their centrifuges, were justified by claiming that nothing else would satisfy the Congressional Joint Committee on Atomic Energy. It was also driven by their desire to see just what progress the British had made since the USAEC ruptured relationships early in 1965. R. Scott Kemp, who has analysed the collaboration with the United Kingdom in the first half of the 1960s using American sources, concludes that the exchanges were, ‘if anything, more beneficial to the United States, which always seemed to be a few months behind’.Footnote 54 By 1969 Britain had, if anything, pulled further ahead. The UKAEA sought to protect their lead, which was also their trump card in their negotiations with the Dutch and the Germans. The USAEC sought to learn what they had done on their own and insisted on having visual access to Mark I. In fact they wanted visual access to all the technology that the UKAEA might share for the next decade with its Continental partners. If originally knowledge flowed freely in the spirit of reciprocity, now it flowed unidirectionally towards the USAEC in an agonistic, competitive relationship, structured by the preponderance of American power.
That power relationship explains why the British authorities decided not to reject outright American demands for visual access (as Zuckerman thought they should). The USAEC, along with the Joint Committee, could impose their will on a majority in Whitehall because the Anglo-American ‘special relationship’ in nuclear matters was asymmetric, with the US at the dominant pole. Senior officials in the British government, notably Denis Healey, anguished over the reprisals that might ensue if the UKAEA did not grant visual access to the centrifuge. An internal memo in November 1968 that surveyed the political risks of having a row with the United States over the gas centrifuge noted the effects that it might have ‘on our bilateral agreement with them for the supply of enriched uranium for our civil programme, on our supplies of this matter for nuclear propulsion and on collaboration on nuclear weapon development, including Polaris’.Footnote 55 A USAEC commissioner confirmed this at the meeting on 3 May 1969. If the British did not grant visual access, US–UK nuclear cooperation in both civil and military fields might be imperilled. The JCAE ‘had always resented’ Britain's refusal ‘to disclose all their R and D work in the course of exchanges with the Americans’, for example in the development of fast breeder reactors. The ‘future of American nuclear co-operation in NATO’ might even be questioned by the Joint Congressional Committee if the US was not given access to the Mark I production model.Footnote 56 In brief, Britain's prototype centrifuge was a knot in the net of civil and military nuclear engagements that were constitutive of the Anglo-American ‘special relationship’, notably immensely important agreements for military cooperation adopted in 1958.Footnote 57 That relationship could begin to unravel if the UKAEA denied access to its centrifuge and even to its pilot enrichment plant – or so it was claimed.
Power: distinguishing the national in the hybrid
A frontier is partially a virtual construction. It is as much a site of the demonstrative extent of power as a real barrier. It regulates an exchange as much as it excludes, but it must be credible enough to deter unlimited flows … The point is that the frontier defines authority, and those who govern lose legitimacy if their frontiers become totally permeable.
Charles MaierFootnote 58Power is rendered visible by the affirmation of boundaries. Article IX (c) of the 1955 US/UK Agreement on Cooperation in the Civil Uses of Atomic Energy specified that neither party could pass material, equipment, devices or restricted data to a third entity without the written consent of the party from which it had been received. And since this was an intergovernmental agreement, enforcing it required identifying the national provenance of ‘pixels’ of knowledge that were embedded in the device, of slicing hybrid, transnationally produced knowledge into nationally identifiable bits and pieces. It involved retroactively reconstructing the national embedded in the transnational. It involved reasserting the authority of the nation state over and against the dilution of its sovereignty implied by reciprocal knowledge flows across a permeable border.
The contest between the USAEC and the UKAEA in 1969 was a contest over the national identity of knowledge. The UKAEA sought to denationalize everything it had learnt from the United States up to 1965, and to nationalize as British all those elements that it wanted to pass on to the Dutch and the Germans – above all, the end caps. The USAEC sought to renationalize every separately identifiable item of information that it had shared with the visitors to Oak Ridge before 1965, and to insist that its identity as American knowledge, while gradually withering away, would survive for another decade or more. Each party to the dispute carved out bits of knowledge from Mark I, and defined them as national and proprietary over and against the hotchpotch of the transnational and the hybrid.
The United Kingdom tried to reinforce the distinction between generally available and specifically exchanged knowledge, between what had been shared in the past and had been developed ‘independently’ since 1965, at every step in these negotiations. The United States tried to dissolve those selfsame boundaries. Mark I, they said, derived from specific American-inspired applications and interpretations of general background knowledge. It was the product of a series of successive micro-choices that were informed by American research and reflected in its design, including decisions not to take certain paths. When the United Kingdom tried to imply that the post-1965 design was their own, developed by combining ‘pure’ British insights with background knowledge that was ‘universal’, the United States insisted that the 1968 centrifuge was a hybrid object that involved the interweaving of ideas from many sources, including data and other insights of American provenance.
Each time the British tried to diminish the national significance of what they might have learnt from the United States, the Americans tried to inflate it. The British argued that the centrifuge that they had in hand in 1969 owed nothing in particular to what they had learnt from the United States when the two had collaborated. There was no ‘American’ knowledge in it. This was because they had decided to orient their work to the development of a prototype (Mark I) that could be used to produce enriched uranium on an industrial scale. This change in objective, from research that was simply intended to learn more, to development of a prototype that could be mass produced, occasioned a discontinuity in the accumulation of knowledge. Mark I drew on what was generally known about this type of device, but the particular features that defined its identity in 1969 were solely British. There were no ‘specific design details’ of American provenance in the UKAEA's centrifuge; thus no ‘specific Restricted Data’ would be revealed to the Dutch and the Germans.
Their American interlocutors emphatically rejected any idea that there was a discontinuity in the historical accumulation of knowledge by the UKAEA as they moved from research to development. They insisted that many items of United States provenance lived on in the United Kingdom's centrifuge, items that could not simply be banished to ‘background technical information which was widely known and unclassified’, items that were specifically American, and so restricted. As Gerald Tape put it in 1969, ‘it was difficult to escape the conclusion that there must be an obvious association between the classified information transferred between 1960 and 1965’ and the current state of the British device.Footnote 59
There is no essence of the national in the hybrid. The national is the negotiated outcome of a quest for origins in the transnational; it is affirmed against the threat of dissolution in the melange. That quest has purpose: the national identity of knowledge is mobilized in a struggle for power. The British were not being deceitful when initially denying that they owed anything in particular to the USAEC; as far as they were concerned the national origins of what they had acquired at Oak Ridge had long since been dissolved. They capitulated after Hill's more careful investigations and Lord Penney's report. For a man like Zuckerman this capitulation was less in recognition of the justice of the commissioner's case than a politically motivated apology needed to placate American demands for inspection and Healey's fear of reprisals. Notwithstanding Penney's report, Sir Solly was emphatic that the end cap was a ‘minor feature in our design that had been discussed at a low-level technical meeting between experts of the two sides’.Footnote 60 This innocuous end cap, an insignificant detail, as he called it, was deemed restricted data by the Americans in a determined effort to get access to the UKAEA's current and future production models, both to evaluate what progress they had made and to see what they could exploit in their own classified centrifuge enrichment programme. As Sir Solly put it to Prime Minister Harold Wilson,
The Americans are out to dominate the world market for nuclear fuel. Were we to allow them access, they might well pick up ideas from our production model which could make a real difference to their commercial exploitation of the centrifuge in third countries, if not in the USA.Footnote 61
Zuckerman saw the demand for access as a cover for industrial espionage.
In 1967 Glenn Seaborg told Walt Rostow, Lyndon Johnson's national security adviser, that the USAEC would continue investigating the technology and economics of gas centrifuge enrichment, notwithstanding the relatively cheap 235U produced by its giant gas diffusion plants. Seaborg wrote that ‘if the process should in the future prove to be an economical method of enriching uranium, the United States should be in a position of world leadership’.Footnote 62 Seen from Washington, the UKAEA's independent progress between 1965 and 1967 seemed to be substantial: why else would they fight tooth and nail to deny visual access? Here was a direct challenge to American scientific and technological leadership. Sir Solly was certainly right to insist that the USAEC was less concerned with security matters than with seeing what progress the UKAEA had made on its own.
Concluding remarks
This paper is framed by James Secord's suggestion that the ‘how’ and the ‘why’ of knowledge circulation is a central question for the history of science, and one that deserves closer attention than heretofore. The ‘how’ was addressed by describing the mundane social processes (Shapin) of knowledge circulation/production that were at work in face-to-face encounters (Raj) between scientists and engineers from the USAEC and the UKAEA who were exploring the properties of gas centrifuges for uranium enrichment. The ‘why’ demanded a multilayered response that engaged ever-widening circles of social actors and their motivations. Reciprocity (Baylis) – the conviction that each had something to learn from the other – brought the two parties together in the first place. It was an a priori condition for the exchange of knowledge between researchers in the two countries. And it only became possible in 1955 in the civil domain, and a bit later in the military field, when the previous American nuclear policy of ‘monopoly and exclusion’ began to thaw.Footnote 63
The collaborative urge enabled by reciprocity was interwoven with a competitive urge driven by national rivalry. Everybody realized that at some stage, depending on the progress made, one or other of the partners might break loose and ‘go it alone’ to protect their proprietary rights. The USAEC chose this path in December 1965. The UKAEA aspired to autonomy in 1969 but was thwarted by the capacity of the American partner to impose its interpretation of the 1955 agreement. Reciprocity was, from its inception, embedded in an asymmetric field of force. What Timothy Botti has said of military cooperation applies here too: ‘Rather than a relationship built on interdependence … as Macmillan had proclaimed in 1957, the Anglo-American nuclear alliance was one in which the United States was the senior party and the British very dependent on the Americans.’Footnote 64
Both partners entered the relationship to learn what they could from the other. Where they differed was in the USAEC's capacity to better lever what they acquired to develop their own programme (if only thanks to superior resources), in their capacity to force open the British research agenda so as to keep an ever-watchful eye on its progress, and in the range of sanctions that they could impose across the full spectrum of nuclear cooperation if they did not get their way. Britain was concerned to secure its leadership vis-à-vis the Dutch and the Germans. The Americans were driven by the quest for ‘world leadership’. Britain aimed to build a European technological community with Continental partners in the civil nuclear domain that effectively sidelined France. America aimed to spread the reach of its civil nuclear imperium across the globe. This quest for leadership (in different registers) is an essential part of the answer to the ‘why’ of knowledge circulation.
The distinction between the ‘how’ and the ‘why’ is analytically useful, but it is dissolved in practice, merging in those who met each other at Oak Ridge, Capenhurst and Risley. As experts in their field they deployed the mundane, culturally sanctioned procedures for collaborating in the co-production of knowledge, as well as the culturally accepted strategies for competitively asserting their proprietary rights. These modes of engagement were also an expression of the national policies of their two governments. While each team collaborated transnationally in order to strengthen national capabilities, the USAEC also sought – and explicitly so from 1969 – to secure global leadership by learning everything they could from an emerging commercial rival and by levering it to their advantage. These policies did not simply circulate in the corridors of Whitehall or within the confines of the Beltway. They were embedded in the face-to-face encounters of the researchers themselves, encounters that were suffused with the asymmetric relationships of power between the two nuclear programmes.
After the four-man visit to Capenhurst and Risley in July 1969 Nelson Sievering of the US State Department, who was in the team, evaluated what had been learnt. He concluded that the British had made a bad design choice. As he put it to Commissioner Thompson,
If the Oak Ridge analysis substantiates the poor potential of the UK production model, especially the limitations which the current end cap imposes, I seriously question whether it's at all in our interest to take steps which would preclude their use of the current end cap designs. It seems to me that the result of denying them the current end cap would be to force them to a better machine than they now have … I cannot help but believe that the sooner we can cut the umbilical cord and make a clean break between the 1960–65 exchanges and the current UK efforts the better off we will be.Footnote 65
The appeal to Article IX (c) had served its purpose. The USAEC had seen what their potential rival was doing, were satisfied that American world leadership was not threatened, and would leave the UKAEA and its partners to blunder on with a suboptimal design for a centrifuge enrichment plant. Transnational collaboration/national rivalry, circulation/denial – these were the parameters that were inscribed in the asymmetric transatlantic network that constituted the American hegemonic enterprise.
