For the abolition of nuclear weapons to be feasible, a necessary condition is international confidence that compliance with an abolition agreement can be ensured. Verification has an important part to play in convincing the leaders of nuclear-armed states and their allies, domestic political opponents, and potentially skeptical publics that the benefits of disarming outweigh the risks. But it cannot accomplish that task by itself. In addition to being able to detect violations reliably and in a timely manner, the international community would also need to be able to act upon them quickly and effectively. Attenuating the security threats that drive nuclear weapons acquisition is yet another part of the challenge. The multifaceted task of building confidence that states intend to remain in compliance with a treaty is a central element of the disarmament challenge. It is important to acknowledge from the outset that the technical approaches discussed in this chapter are vital, but still only one part of the overall challenge.
Two distinct phases can be discerned in the task of verifying nuclear disarmament.1 During the transition to zero, states would want confidence that all nuclear weapons were being eliminated and all fissile material was being placed under appropriate international safeguards. Once this had been accomplished, ongoing verification of all remaining nuclear activities would be necessary to detect any attempts at rearmament.
Ironically, we probably have a clearer conception of the ongoing verification that would be needed to detect rearmament in a nuclear weapons–free world than of the scheme required to verify the elimination of nuclear weapons in the first place. Safeguards against rearmament would involve verifying the nondiversion of nuclear material in declared civilian facilities, and also that no state was conducting undeclared nuclear activities. This is essentially the task performed by the International Atomic Energy Agency (IAEA) today—at least in non–nuclear weapon states that have an Additional Protocol in force.2 In a nuclear weapon–free world, however, such safeguards would presumably need to be applied to all states, including the former nuclear-armed states. Moreover, before dismantling their last nuclear weapons, the nuclear-armed states would also probably need to be convinced that these safeguards were considerably more effective than they are today. For this reason, some have suggested that traditional safeguards may need to be augmented by eliminating the most proliferation-sensitive nuclear activities or by placing them under some kind of international or multinational control. Nevertheless, there has been extensive research into safeguards for civilian nuclear activities, and a large body of experience has already been built up in applying them.
Verifying the transition to zero presents more of a challenge. As with all arms-reduction treaties, a key task would be to verify that declared weapons were being dismantled as claimed. Where nuclear weapons are concerned, this process would be complicated by the existence of classification rules that prevent international inspectors from viewing warheads directly or using most, if not all, standard measurement techniques for assaying fissile material. Largely because of this, no previous nuclear-arms control treaty has placed verified limits on warheads. Nevertheless, in expectation of future arms-control treaties that do concern warheads, there has been extensive research into the problem.3 Originally, this research was confined to U.S. and Russian national labs, but it is now being conducted more broadly. Although there are a number of technical challenges still to be overcome, it seems likely that, with enough effort, adequate solutions can be found.
Ultimately the toughest verification challenge would be to ensure that no nuclear-armed state has retained any undeclared stockpiles of fissile material, whether in the form of nuclear weapons or not. It involves “proving a negative”—that is, verifying the absence of something.4 The challenge with this, of course, is that it is impossible to check each and every conceivable hiding place for illicit material, and there is always the possibility that something might have been stashed in a place that hasn’t been inspected. In the case of fissile material the challenge is exacerbated by the ease of shielding its radioactive emissions, making it undetectable at anything more than a few meters.
Fortunately, there is a solution to this problem—at least at a conceptual level. In theory, the inspectorate could uncover undeclared material by first verifying a state’s declaration of the quantity of fissile material it had produced and used. Then inspectors would compare the amount of material that a state ought to have (its estimated inventory) to the amount it actually did have (its measured holdings). If a state had declared all its fissile material to inspectors, these two quantities should be equal. If, on the other hand, a state had secreted some nuclear material away, its estimated inventory would be larger than its current holdings. In practice, there are inevitably uncertainties associated with estimating the quantity of fissile material produced and used by a state and with measuring its current holdings. Consequently, the amount of material that ought to be present would never be exactly equal to the amount actually measured, and there would be no way of knowing whether any discrepancy was due to measurement errors or diversion. If the difference were smaller than the amount needed to produce a nuclear weapon, this might not matter much. However, if the difference were larger (as the following analysis shows it would be), the international community might worry that the state in question had secreted away a militarily significant quantity of fissile material—and this could prove a barrier to abolition.
The first step in the verification process would be for the nuclear-armed states to submit comprehensive declarations detailing the complete history of their fissile material production programs for each weapons-usable fissile material (that is, not just uranium-235 and plutonium-239, but also uranium-233, neptunium-237, and possibly others). The declarations would include detailed information on current holdings of these materials.
The inspectorate might be able to build some confidence in the veracity of information on past production by checking the records for internal consistency, as well by as subjecting them to traditional forensic analysis (that is, checking that the paper and ink are of the right age, and so forth). Such analysis has its limitations, however. Records are often incomplete or erroneous.5 Moreover, the move from paper to computer records in the past decade or two has had the side effect of making it significantly easier to doctor them. States’ records, by themselves, would be insufficient for the inspectorate to verify the veracity of their declarations.
“Nuclear forensics” provides an independent means of reconstructing fissile material production by examining the physical evidence left by the production techniques.6 For instance, the majority of plutonium for weapons programs was produced in so-called graphite-moderated reactors. All graphite contains very small quantities of impurities. During reactor operation, some of these impurities capture neutrons to form isotopes that are not naturally occurring. By measuring the abundance of these “activation products” it is possible to estimate the total quantity of plutonium produced in the reactor.7 This method is known as GIRM—the graphite isotope ratio method.
GIRM has the potential to reduce the uncertainty in the amount of plutonium produced in a reactor significantly—to about 5 percent, in fact. Moreover, it has been validated successfully in a field trial.8 In 1995–97 it was applied to the Magnox reactor at Trawsfynydd in the UK, which was then being decommissioned. It was estimated that the total quantity of plutonium produced in this reactor was 3.63±0.19 MT—an error of about 5 percent. This is a formidable technical achievement. Nevertheless, even this uncertainty represents sufficient material for about twenty nuclear weapons in a state with some experience of weapons design. Moreover, GIRM does nothing to reduce any uncertainty in the amount of plutonium that a state actually separated from the spent fuel.
Graphite-moderated reactors are, however, not the only means for producing fissile material. Heavy water–moderated reactors account for the remaining eleven out of forty-five reactors that have been used to produce plutonium for weapons programs. Moreover, the vast majority of weapons-origin fissile material is high enriched uranium. No equivalent forensic techniques exist for estimating fissile material production in either heavy water–moderated reactors or enrichment facilities. Even if such techniques were developed it would not be possible to apply them in all cases because a number of enrichment plants are being, or have been, disassembled.
Although not as severe, there are also uncertainties associated with measuring a state’s holdings of fissile material. Any measurement is accompanied by some uncertainty. Where large quantities of nuclear material are concerned, the cumulative effect of uncertainties on multiple measurements can be sig nificant, leading to a noticeable difference between the amount of fissile material that is calculated to be present and the amount that is actually measured. In IAEA terminology this difference is known as material unaccounted for (MUF). As mentioned above, the problem is that it can be impossible to determine whether MUF is due to measurement errors or the diversion of nuclear material.
The existence of MUF is a source of controversy leading some to question the effectiveness of today’s IAEA safeguards system.9 Whether or not criticisms are valid, MUF is inevitable. As British Nuclear Fuels, for instance, pointed out in a press release (following some media controversy about MUF at Sellafield):
These uncertainties exist in all industrial processes, for example the gold industry experiences the same thing when extracting gold from ores—the amount recovered never precisely matches the amount estimated in the ore.
No nuclear material has been stolen. Figures change from year to year. Negative numbers do not mean material has disappeared; positive numbers don’t mean material has been created.10
An idea of the likely magnitude of the MUF associated with nuclear weapons programs can be found in recent UK and United States efforts to account for the material in their fissile material production programs.11 Both states calculated how much plutonium and HEU they ought to have (shown in row 1 of Table 13.1) and measured the amounts they actually did have (shown in row 2). The MUF (the difference between these two figures) is shown in row 3. In the case of UK it is negative because the UK found more material than it thought would be present. In the case of the United States, the MUF would be enough for about 1,000 warheads.
In fact, from the perspective of international inspectors, there would be an even bigger source of uncertainty. Some of the material in the UK and United States inventories would be unavailable for verification. For instance, it would be impossible to accurately verify the quantity of material used in nuclear tests—even if the state that conducted the tests knew those quantities accurately. Row 4 of the table shows a lower bound for the amount of “material unavailable for verification” (MUV) based on the quantities estimated to have been used in tests. In fact, the MUV would probably be much larger because some of the material in the UK and United States inventories was burned in reactors, lost in waste streams or decayed—and, again, verifying such losses to any accuracy would prove impossible.
Calculated and Measured Quantities of Fissile Material in the UK and U.S. Military Stockpiles
SOURCE: This table originally appeared in George Perkovich and James M. Acton, Abolishing Nuclear Weapons, Adelphi Paper 396 (London: Taylor and Francis, Ltd., 2008), ch. 2, reprinted by permission of the publisher.
These various sources of error—both in estimating past production and in measuring current holdings—would make it impossible for an inspectorate to definitively prove that a disarming state had not retained a clandestine stockpile of fissile material containing a few percent of its total holdings. In the case of the United States or Russia this would be sufficient for literally thousands of warheads.
Although this conclusion seems (and indeed is) rather pessimistic, it does not necessarily imply that nuclear disarmament is infeasible. In fact, on one previous occasion the IAEA did succeed in convincing the international community that a state had completely eliminated a homemade nuclear arsenal and placed all of its fissile material under international safeguards. The state in question was South Africa, and the lessons from verifying the elimination of its nuclear weapons program are potentially instructive.12
Starting in 1979, South Africa manufactured six nuclear weapons and an uncompleted seventh. All of these devices used high enriched uranium. In 1990 South Africa made the decision to disarm. It dismantled these weapons and, in 1991, acceded to the Nuclear Non-proliferation Treaty (NPT) and submitted its “initial declaration” to the IAEA. This document detailed its holding of fissile material.
For all the reasons discussed above, verifying the completeness of this declaration proved a significant challenge. The results of “technical” verification activities were inconclusive; that is, the IAEA could not rule out the existence of a clandestine stockpile of HEU sufficient for one or more nuclear weapons. The strongest conclusion that the IAEA director-general was able to reach was that “having regard to the uncertainties normally associated with data of this nature, it is reasonable to conclude that the uranium-235 balance . . . of the pilot plant is consistent with uranium feed.”13
Nevertheless, that uncertainty did not lead to states questioning South Africa’s good faith or intentions. Part of this success (but by no means all of it) was due to South Africa’s open and transparent behavior, beyond any legal requirement: it provided inspectors with a comprehensive history of its nuclear program, gave them all relevant records, and provided them with access to all facilities, materials, and personnel. It cooperated fully to resolve any discrepancies that did arise. This transparency did not reveal information that allowed the IAEA to definitively rule out the existence of clandestine HEU; rather, it convinced the inspectors on the ground, as well as outside observers, that South Africa had nothing to hide. South Africa provides a clear example of how transparency can help substitute for unavoidable deficiencies in technical verification.
Applying this model to today’s nuclear-armed states would certainly be a tremendous challenge and, it is necessary to acknowledge, may not be possible. They have produced much more fissile material than South Africa. More important, they could have much more reason to cheat. The goodwill shown to South Africa after the end of apartheid cannot be ignored either. But, equally, it would be wrong to dismiss the relevance of the South Africa model. What should be tried is more than just a “one off” verification attempt in which the nuclear-armed states are submitted to a single, intensive round of verification. Rather, a slower and more gradual process should be envisaged in which the nuclear-armed states gradually become more transparent about their fissile material and are proactive in finding ways to build confidence in the veracity of their claims. Over time, if no irresolvable evidence emerges to contradict these claims, states might be much more willing to tolerate uncertainties associated with verifying fissile material.
One important element in such a confidence-building process would be a verified fissile material cut-off treaty (FMCT), including stocks. Such a treaty would involve the nuclear-armed states declaring existing stocks of fissile material and submitting them—presumably as they became excess—to some kind of verification. Because the nuclear-armed states would still retain their nuclear weapons as a hedge, a fissile material cut-off treaty would require less stringent verification than a treaty eliminating nuclear weapons. However, if stocks were successfully verified (and it may be a gradual process that is only possible as stocks are declared excess) and if, in the following years, no irresolvable evidence came to light that any state had failed to declare all of its stocks, the international community would be presented with a much more promising “baseline” from which to contemplate the complete elimination of nuclear weapons. That said, the prospects for including stocks in a fissile material cut-off treaty are very poor.14 An FMCT treaty that did not include them would still be a worthwhile and important step—but not nearly as significant if elimination is the ultimate goal.
Unfortunately, even though negotiations on an FMCT in the Conference on Disarmament have now commenced, a fissile material cut-off treaty cannot be considered a short-term goal. The remainder of this chapter sketches out the first steps that could be taken by the nuclear-armed states in a confidence-building process. Non–nuclear weapon states should be willing to publicly recognize them as meaningful steps toward zero.
In the short term it is necessary to recognize that greater transparency about capabilities by China, India, Israel, and Pakistan could be detrimental to disarmament efforts by stimulating a competitive reaction from their adversaries.15 A confidence-building process must be tempered by this concern while working to build relations that are conducive to greater transparency.
As mentioned above, one of the most significant challenges to be faced in the long term is that states’ own records of their nuclear programs are incomplete, either because they were never kept in the first place (especially in the early days of nuclear weapon programs) or because they have been lost or destroyed. Records are important, not just as evidence to support declarations, but also because making them available to inspectors is a powerful sign of good faith (à la South Africa). Conversely, the failure to produce the records requested by inspectors could be interpreted as an indication that a state had something to hide, even if it were genuinely because the state was unable to produce the records in question.
The failure of states to keep comprehensive records, along with the deaths of the first generation of personnel associated with many nuclear weapon programs, means that some information that would be useful for verification has been irretrievably lost. However, the nuclear-armed states should take action to avoid this situation deteriorating further. Specifically, they could appoint national commissions to compile comprehensive records of their nuclear programs by, among other steps, collecting documents and interviewing personnel. These records would not have to be made public in the first instance. States could reveal as much or as little of them as they felt able to. The crucial point is that by collecting this information—even if it were kept secret for the time being—states would be facilitating more robust verification further down the line. For this reason, the value of the exercise would be increased even further if the nuclear-armed states were to agree among themselves what information would be useful, as well as relevant standards for record keeping.
One objection to compiling nuclear histories (and indeed, some of the other verification measures discussed in this paper) is expense. The specific justification for compiling nuclear histories is that they provide a baseline against which measurement of current holdings can be measured and, further, that they are a very powerful confidence-building measure and hence provide value for money. More generally, however, there are various points to be made when analyzing the costs of verification. First, on the scale of nuclear weapon budgets, the costs of any verification activities are very modest. Second, the short-term costs of verification need to be compared against the long-term savings of reducing and eventually eliminating nuclear weapons. Third, many of the verification measures discussed in this paper, being directed toward improved accounting of fissile materials, also serve to prevent nuclear terrorism. They are therefore worth doing irrespective of disarmament goals.
As described above, both the UK and the United States have published some details of their fissile material production histories and current holdings. To date, none of the other nuclear-armed states have done so—but they could and should follow suit. In general, the more information that was included, the more useful the declarations would become. In this regard, there is certainly scope for the UK and the United States to supplement their declarations. For instance, the UK gives quantities of HEU in terms of tonnes of uranium (thereby not distinguishing between material of different enrichment levels), whereas the U.S. declaration uses the somewhat more useful unit, tonnes of U-235. The UK could follow the U.S. lead in this regard, and both states could consider giving more details about the various enrichment levels of their HEU holdings. Nevertheless, it is important to point out that, in spite of these areas for further development, the UK and the United States are still ahead of the other nuclear-armed states merely by having published any data at all.
Civilian plutonium and high enriched uranium are as important from the perspective of disarmament as military fissile material. After all, they are all weapons-usable (even if civilian plutonium is typically of a higher burn-up than military plutonium and hence less suitable for military ends).
One positive development in this regard is that since 1998, the five nuclear-weapon states (along with Belgium, Germany, Switzerland, and Japan) have published details of their civil plutonium holdings in so-called INFCIRC/549 declarations, as part of a larger IAEA project on managing civilian plutonium.16 France and the UK (along with Germany) include civilian HEU holdings in these declarations. The other weapon states could emulate this. Moreover, India, Israel, and Pakistan could consider joining this initiative, which, because it relates to civilian and not military material, might be feasible.
This laudable initiative would be enhanced by placing more civilian nuclear material in the nuclear-weapon states under IAEA safeguards—a step that would help demonstrate the veracity of these declarations. All of the nuclear-weapon states have “voluntary offer agreements” with the IAEA.17 These specify both the aim and scope of safeguards (which vary from state to state) and include a (confidential) list of the facilities that the IAEA is permitted to safeguard. These offers vary greatly in their comprehensiveness. At one extreme are the British and French offers, which apparently include every civilian facility (these are already subject to Euratom safeguards anyway); at the other extreme are the much more limited Chinese and Russian offers. A useful first step would be for China, Russia, and the United States to expand the scope of their voluntary offers.
In practice, because of severe budgetary constraints, the IAEA only safeguards a very small number of these facilities. Indeed, there is an ongoing discussion about whether three new gas centrifuge plants to be built in the United States will be safeguarded; the IAEA is reluctant to do so because of the cost. The nuclear-weapon states could consider, therefore, paying for the IAEA to safeguard more of their facilities. They should be prepared to meet the full costs of safeguards, including the costs of employing the extra inspectors required so that scarce inspection resources are not diverted from non–nuclear weapon states. The goal should be to place all civilian nuclear activities in the weapon states under safeguards, even if it may take years to accomplish.
A more comprehensive, multilateral framework for dealing with fissile material stocks, the Fissile Material Control Initiative (FMCI), has been proposed by special adviser to the U.S. Department of State Robert J. Einhorn. He describes it as follows:
FMCI would be a voluntary, multilateral arrangement open to any country that possessed fissile material (whether safeguarded or not) and was willing to sign onto a set of agreed principles. The overall goals of FMCI would be to increase security, transparency, and control over fissile material stocks worldwide; to prevent their theft or diversion to non-state actors or additional states; and to move fissile materials verifiably and irreversibly out of nuclear weapons and into forms unusable for nuclear weapons.
FMCI would establish an agreed set of guidelines that partners, as appropriate to their particular fissile material holdings, would be encouraged to follow. The guidelines would call on the partners:
• to make regular declarations regarding their fissile material stocks by category;
• to apply the highest standards of physical protection and accountancy to those stocks;
• to declare regularly amounts of material they regard as excess to their weapons needs;
• to place such excess material under IAEA safeguards as soon as practicable; and
• to convert excess material as soon as possible to forms that cannot be used for nuclear weapons (e.g., by blending down HEU to low-enriched reactor fuel).18
Einhorn’s premise is that it is very unlikely that stocks will be included in an FMCT because of the opposition of most of the nuclear-armed states. Accordingly, he argues that FMCI could make agreement on an FMCT more likely by helping to address the concerns of states that wished stocks to be included in an FMCT. I believe that this is correct and, moreover, that FMCI would also increase the likelihood of achieving an FMCT that included stocks by getting nuclear-armed states used to the type of verification and transparency measures that would be needed.
This chapter is intended to give an overview of the challenges associated with the most difficult aspect of verifying the elimination of nuclear weapons—accounting for fissile material—and of how these challenges might be overcome. In the final analysis, however, the issue of whether states would be willing to accept the inevitable uncertainties of verification is an essentially and unavoidably political question. For this reason, this paper will close with a very brief exploration of the political question: how effective would verification need to be for the nuclear-armed states to disarm completely?19
The “classical” or Wiesner model of verification argues that as a state reduces the size of its nuclear arsenal in accordance with treaty commitments, the size of a violation by an adversary that would upset the strategic balance also decreases.20 For instance, when the United States went about verifying the 1987 Intermediate Nuclear Forces Treaty it put in enough inspection effort to assure itself that the Soviet Union had retained no more than about fifty SS-20 missiles (even though the agreement required the Soviet Union to eliminate that type of weapon completely), because it deemed that a violation involving fewer than fifty missiles did not pose a military threat to the United States.21 Had the United States possessed a smaller arsenal when the treaty was concluded, the size of a violation that it deemed to be militarily significant would have been smaller and correspondingly more inspection effort would have been required. This Wiesner model predicts that even a very small violation would be militarily significant in a nuclear weapon–free world, and, in consequence, almost perfect verification would be needed. If this model is correct, the prospects for disarmament are dim.
However, there has been little, if any, serious enquiry into the validity of the Wiesner model or its assumptions. At least three questions can be asked of this model:
• In a nuclear weapon–free world would the size of a militarily significant violation really be so small as to make effective verification infeasible? Or, in plainer terms, would the retention or creation of a very small nuclear arsenal actually upset the strategic balance?
• The Wiesner model implicitly assumes that the level of distrust between states remains unchanged during the verification process. But, would the implementation of a verification scheme that produced no irresolvable evidence of cheating build states’ confidence in each other’s intentions and hence prevent spiraling demands for verification?
• Even if verification were perfect, effective enforcement would still be required. Could improvements in enforcement mechanisms therefore compensate for verification deficiencies?
There are no easy or obvious answers to any of these questions, but addressing them would help to assess how good verification needs to be. They therefore deserve as much attention as the mechanics of verification.
Some of the material in this chapter draws upon George Perkovich and James M. Acton, Abolishing Nuclear Weapons, Adelphi Paper 396, ch. 2 (London: Taylor and Francis, Ltd., 2008), http://www.informaworld.com.
1. For a detailed discussion of verifying disarmament, see George Perkovich and James M. Acton, Abolishing Nuclear Weapons, Adelphi Paper 396 (London: Taylor and Francis, Ltd., 2008), chs. 2 and 3. This volume is available in George Perkovich and James M. Acton, eds., Abolishing Nuclear Weapons: A Debate (Washington, DC: Carnegie Endowment for International Peace, 2009), section 1, http://www.carnegieendowment.org/ files/abolishing_nuclear_weapons_debate.pdf. See also Steve Fetter, “Verifying Nuclear Disarmament,” Henry L. Stimson Center Occasional Paper 29, October 1996, http://www.stimson.org/wmd/pdf/fetter.pdf; Committee on International Security and Arms Control, National Academy of Sciences, Monitoring Nuclear Weapons and Nuclear Explosive Materials: An Assessment of Methods and Capabilities (Washington, DC: National Academies Press, 2005), at http://www.nap.edu/catalog.php?record_id=11265.
2. Contrary to popular belief, the IAEA’s legal authority to verify the absence of undeclared nuclear material does not stem from the Additional Protocol, but from Article 2 of the Comprehensive Safeguards Agreement, which asserts “the Agency’s right and obligation to ensure that safeguards will be applied . . . on all source or special fissionable material” (my italics) (IAEA, “The Structure and Content of Agreements between the Agency and States Required in Connection with the Treaty on the Non-Proliferation of Nuclear Weapons,” INFCIRC/153 [Corrected], June 1972, http://www.iaea.or.at/Publications/Documents/Infcircs/Others/infcirc153.pdf). The Additional Protocol does, however, give the IAEA the practical tools it needs to be able to draw credible conclusions about the absence of undeclared nuclear material.
3. Office of Nonproliferation Research and Engineering, “Technology R&D for Arms Control,” Arms Control and Nonproliferation Technologies (Spring 2001): 4–17, http://www.fas.org/sgp/othergov/doe/acnt/2001.pdf; Committee on International Security and Arms Control, National Academy of Sciences, Monitoring Nuclear Weapons and Nuclear Explosive Materials, pp. 97–108.
4. For a discussion of the problem of “proving a negative,” see John Carlson, “Safeguards in a Broader Policy Perspective: Verifying Treaty Compliance,” paper presented to the conference on “Changing the Safeguards Culture,” Santa Fe, NM (30 October–2 November 2005), www.asno.dfat.gov.au/publications/2005_santa_fe_policy.pdf.
5. See, for example, Thomas W. Wood, Bruce D. Reid, John L. Smoot, and James L. Fuller, “Establishing Confident Accounting for Russian Weapons Plutonium,”Nonproliferation Review 9, no. 2 (Summer 2002): 134, http://cns.miis.edu/pubs/npr/vol09/92/ 92wood.pdf.
6. Steve Fetter, “Nuclear Archaeology: Verifying Declarations of Fissile-Material Production,” Science and Global Security 3, nos. 3–4 (1993): 240–44, http://www.princeton.edu/sgs/publications/sgs/pdf/3_3-4Fetter.pdf.
7. In practice the measured quantity is actually the mass ratio between impurities and their activation productions.
8. Wood et al., “Establishing Confident Accounting for Russian Weapons Plutonium,” pp. 130–31.
9. Henry Sokolski, ed., Falling Behind: International Scrutiny of the Peaceful Atom (Carlisle, PA: Strategic Studies Institute, February 2008), http://www.npec-web.org/ Books/20080327-FallingBehind.pdf, ch. 5.
10. British Nuclear Fuels PLC, “Media Response: Publication of Materials Unaccounted For,” 13 August 2005, http://www.bnfl.com/content.php?pageID=49&newsID=70.
11. U.S. Department of Energy, “Plutonium: The First 50 Years. United States Plutonium Production, Acquisition, and Utilization from 1944 through 1994,” DOE/DP-0137, February 1996, http://www.fissilematerials.org/ipfm/site_down/doe96.pdf; U.S. Department of Energy, “Highly Enriched Uranium: Striking a Balance: A Historical Report on the United States Highly Enriched Uranium Production, Acquisition, and Utilization Activities from 1945 through September 30, 1996,” draft, revision 1, January 2001, http://www.fissilematerials.org/ipfm/site_down/doe01.pdf; UK Ministry of Defence, “Historical Accounting for UK Defence Highly Enriched Uranium,” March 2006, http://www.fissilematerials.org/ipfm/site_down/mod06.pdf; UK Ministry of Defence, “Plutonium and Aldermaston: An Historical Account,” 2000, http://www.fas.org/news/uk/000414-uk2.htm.
12. For an account of denuclearization in South Africa, see Adolf von Baeckmann, Gary Dillon, and Demetrius Perricos, “Nuclear Verification in South Africa,” IAEA Bulletin 37, no. 1 (1995), http://www.iaea.org/Publications/Magazines/Bulletin/ Bull371/37105394248.pdf; and Darryl Howlett and John Simpson, “Nuclearisation and Denuclearisation in South Africa,” Survival 35, no. 3 (Autumn 1993).
13. IAEA, “The Agency’s Verification Activities in South Africa,” GOV/2684 (8 September 1993), para. 29.
14. For a discussion of issues confronting the FMCT, see Jenni Rissanen, “Time for a Fissban—or Farewell?” Disarmament Diplomacy no. 83 (Winter 2006), http://www. acronym.org.uk/dd/dd83/83fissban.htm.
15. Perkovich and Acton, “Abolishing Nuclear Weapons,” pp. 40–42.
16. “Guidelines for the Management of Plutonium (INFCIRC/549): Background and Declarations,” Institute for Science and International Security (1 April 2004, revised 16 August 2005), http://isis-online.org/global_stocks/end2003/infcirc_549.pdf.
17. For a comprehensive description of IAEA safeguards in the nuclear weapon states, see International Panel on Fissile Materials, Global Fissile Materials Report 2007 (International Panel on Fissile Materials, 2007), ch. 6, http://www.fissilematerials.org/ ipfm/site_down/gfmr07.pdf.
18. Robert J. Einhorn, “Controlling Fissile Materials and Ending Nuclear Testing,” paper presented to the conference on “Achieving the Vision of a World Free of Nuclear Weapons: International Conference on Nuclear Disarmament,” Oslo (26–27 March 2008), http://disarmament.nrpa.no/wp-content/uploads/2008/02/Paper_Einhorn.pdf. See also Ivo Daalder and Jan Lodal, “The Logic of Zero: Toward a World without Nuclear Weapons,”Foreign Affairs 87, no. 6 (November–December 2008): 86–90.
19. For a somewhat lengthier treatment of this issue, see Perkovich and Acton, “Abolishing Nuclear Weapons,” pp. 49–52.
20. Allan S. Krass, Verification: How Much Is Enough? (London: Taylor and Francis for SIPRI, 1985), pp. 167–71.
21. Office of Technology Assessment, U.S. Congress, Verification Technologies: Cooperative Aerial Surveillance in International Agreements (Washington, DC: U.S. Government Printing Office, July 1991), p. 104, Box C-1, n. 1, http://www.princeton.edu/~ota/ disk1/1991/9114/9114.PDF.