The American Journal of International Law
July, 1987, 81 A.J.I.L. 593
South Asia is now poised for a nuclear arms race. Pakistan has learned how to make enriched uranium — the material that destroyed Hiroshima — and has been buying the electronic switches and hollow steel spheres used for implosion. It has tested, successfully, an implosion bomb with a dummy core. On the Indian side, Prime Minister Rajiv Gandhi has been saying that India could make an atomic bomb “in a matter of several weeks” and “could have done so for the past ten or eleven years.”
These developments pose two risks. The immediate one is that Pakistan will conduct an explosive test, which would force an Indian deployment, or that the Indians will deploy, which would force a Pakistani test. Under current U.S. law, a Pakistani test ends U.S. aid and effectively ejects the United States from Pakistan. If this happened, South Asia would be left with a pair of nuclear-armed rivals face to face, and the Soviets still in Afghanistan.
The second risk is longer term, but possibly greater: the international controls on nuclear exports may simply collapse after such a failure in South Asia. Why? Because India and Pakistan have built their programs from imports. In nuclear South Asia, nothing is “indigenous.” Almost every essential facility in India and Pakistan has been imported directly, copied from imports or built with foreign designs. If these two countries deploy atomic bombs, it will be obvious that the export control system has not worked.
It is India that dominates the South Asian arms race. India’s greater wealth, population and conventional strength mean that Pakistan must either match India in nuclear arms or be dwarfed. Pakistan has now struggled up to the ability to test; therefore, India must decide the next step. It could be to build an arsenal. If India does, it will be by using exports from the United States, Canada and the Soviet Union.
This poses a serious issue for all these countries, and especially the United States. So far, the United States has been India’s largest nuclear supplier. Controls attached to U.S. exports require the peaceful use of power reactors, research reactors and nuclear fuel. In the past, the controls have not been used wisely; at present they are hardly used at all. But it remains true that if the controls were used vigorously, and in league with India’s other nuclear suppliers, there is still a chance that India’s program could be restricted to peaceful ends. If this were achieved, the next stage of escalation in South Asia would be headed off, and the world’s struggle against nuclear arms would get a giant boost.
United States control comes from reactors and the materials needed to run them. As these words are written, uranium is being enriched under the U.S.-India agreement covering the Tarapur reactors. That uranium will be shipped to India by 1987 or 1988, made into reactor fuel by 1989, burned in and removed from reactors by 1992, and cooled in a reactor storage pool until about 1993. Then, according to India, the spent fuel’s plutonium will be legally free for atomic bombs. Plutonium, like high-enriched uranium, is a nuclear explosive: 6 kilograms of it destroyed Nagasaki in 1945. It is created whenever uranium fuel is irradiated in a reactor. India contends that in 1993, just 6 years from now, the two reactors at Tarapur and all the plutonium in their spent fuel will be free for use in bombs. India relies on an implausible and stubborn reading of the expiration clause of the Tarapur agreement. This reading, which will be discussed below, would give India about 1,800 kilograms of completely unrestricted plutonium, enough for 225 atomic bombs. By using only the “weapon grade” plutonium in this stockpile, India could make about 36 bombs.
If India gets its way, it will mean the following: first, that plutonium made from all the enriched uranium now being supplied, and all the enriched uranium to be supplied under the Tarapur agreement will be free for use in bombs as soon as the plutonium is physically ready; second, that the U.S.-supplied Tarapur reactors will have functioned as military production reactors for their entire lives, if India so chooses; third, India will have become the first country in the world to declare a peaceful U.S. nuclear export program free for military use, and will have done so openly and, according to India, legally; and fourth, India will have torn a precedential hole in the legal fabric surrounding U.S. nuclear exports, because most U.S. agreements have the same language on expiration as India’s. In general, India will have shown that the United States cannot control its nuclear exports. If India shows that, the U.S. effort to restrain other nuclear suppliers will be set back, and India will have struck a second blow in favor of nuclear arms proliferation. The first was in 1974, when India became the first and only country in the world to explode an atomic bomb made from “peaceful” nuclear imports.
How did the United States get into such a position? And what can be done about it?
INDIA’S NUCLEAR PROGRAM
India wants two things from its nuclear program: electricity, and the option to make atomic bombs. The second is as important as the first. In 1976 India accepted an expensive delay in getting its second reactor built at Rajasthan, rather than bow to Canada’s demand for peaceful use of other Indian facilities. And in the early 1980s, India let a fully completed reactor in Madras remain idle for 2 years, because importing the heavy water to run it would have meant restricting it to peaceful use.
But progress toward arms has been slow. Until 1986, supplier controls stood in the way. India started its first important reactor, called CIRUS, in 1964. Canada supplied the reactor, and the United States the “heavy water” to run it. Heavy water (deuterium oxide) is needed to achieve a chain reaction with the natural uranium fuel that CIRUS uses. India promised Canada and the United States to restrict CIRUS to peaceful use, but CIRUS was not covered by international inspection; consequently, in 1974 India could test a bomb made with plutonium from CIRUS’s spent fuel and call it a “peaceful nuclear device.”
After the explosion, the Department of State was asked for explanations. It said that India had “commingled” U.S.-supplied heavy water with domestic Indian heavy water, and that the commingled water was used in CIRUS. But State admitted that U.S. heavy water was the only water in CIRUS when CIRUS was started up, and that U.S. heavy water must have been in CIRUS when it made the plutonium for the test. Nevertheless, State claimed that there was no way to establish the origin of the plutonium “conclusively.” Although Congress was outraged, State did nothing about the explosion and did not insist on any system for keeping track of U.S. heavy water. The result is that the United States still has a clear right to peaceful use of at least some of CIRUS’s plutonium — tha is, of a fraction equal to the fraction of U.S. heavy water in the reactor. Given the way CIRUS operates, the fraction could be large. Heavy-water losses from a reactor such as CIRUS amount to about 0.3 percent per year, which would add up to a 1-tonne loss out of a 19-tonne inventory over 23 years of operation. Thus, unless India deliberately took the U.S. water out, all except 1/19 of CIRUS’s water is still of U.S. origin and the United States has the right to peaceful use of all except about 1/19 of the CIRUS plutonium.
India also gave Canada a guarantee of peaceful use for the CIRUS reactor. Canada ended its nuclear cooperation with India because of the explosion in 1974, but that did not put Canada in breach of its CIRUS agreement or excuse India from performing it. In view of India’s action, Canada was entitled to suspend its performance until India promised not to make nuclear weapons. Since India has not given this promise, Canada’s performance remains suspended while India’s duties remain.
In short, some and probably most of the plutonium produced by CIRUS is covered by a peaceful use pledge to the United States, and all of it is covered by a similar pledge to Canada. However, as the plutonium is not under international inspection, there is no official way of knowing what it is being used for.
For its next two sets of reactors, India had to agree to international inspection — called safeguards — as well as peaceful use of all the plutonium they would make. These promises covered the twin Tarapur reactors, built by the United States in the late 1960s, and the twin Rajasthan reactors, built by Canada in the 1970s and early 1980s. It was not until 1983, with the first reactor at Madras, that India began to escape controls. India itself had built Madras by copying Canada’s design for Rajasthan. But because “replication” was not forbidden by the India-Canada agreement, India could say that Madras was indigenous, and its plutonium unrestricted.
There was still, however, a hitch. Madras, like Rajasthan, needed heavy water to operate. And for it to operate free of safeguards and pledges of peaceful use, India itself would have had to produce the heavy water. India could not make enough on its own — it was short a vast amount — hence India was still not free. There were only four suppliers of heavy water in the world: Canada, China, the United States and the Soviet Union. All except China required peaceful use and international inspection of any plutonium made with their nuclear exports. There is very strong evidence that in 1983, when Madras was started up, India either imported 100 metric tons of heavy water from China — secretly — or diverted that amount from international inspection. India had to do one or the other to fill the gap in its own production. Whichever it did, Madras was started free of controls, which made its plutonium available for bombs.
India began to extract the Madras plutonium in 1986. Extracting plutonium from spent fuel requires a special factory with heavy shielding, and takes several months. The plutonium emerges in pure form, and can be used to fuel additional reactors or to make atomic bombs. Through 1986, India could have extracted about six bombs’ worth — not quite enough for an arsenal.
Soon there will be enough. In 1985 India started a second reactor at Madras — another copy of Rajasthan — and a second research reactor called Dhruva — a scale-up of CIRUS. Neither is restricted. Together with the first reactor at Madras, they make enough plutonium for about 15 bombs per year. But they all require heavy water. India’s heavy-water shortage in 1985 was even worse than in 1983; therefore, these reactors too must be using the Chinese import, or heavy water diverted from safeguards.
Finally, India operates the Fast Breeder Test Reactor, built with help from France. It is fueled by plutonium from CIRUS; thus, any plutonium it makes should be covered by CIRUS’s peaceful use restriction. The Canadian and U.S. agreements under which India got CIRUS exclude “use” for any military purpose. If CIRUS is used to make plutonium cores for the Breeder, and those cores are used to breed plutonium for a weapon, then CIRUS will in fact have been “used” for a military purpose. There is a direct causal chain. Without CIRUS, there would be no weapon, regardless of whether one made a weapon from CIRUS’s product directly or by using the product in additional manufacturing.
An example may make this clearer. If the United States supplied a country with a valuable organism for biological research in which the organism would be used to create others, both the first generation of other organisms created and the subsequent ones would have been created by the “use” of that organism. At the time of the CIRUS agreement, it seemed clear that the peaceful use guarantee would reach all generations of fissile material CIRUS would make. The guarantee would not have much meaning otherwise. In ordinary understanding, peaceful use means that CIRUS will not be used in a weapons-manufacturing process, regardless of whether CIRUS is one or two steps away from the final product. The Breeder just started, and will make slightly more plutonium than it uses.
What, then, does one conclude about the Madras reactors, Dhruva and the Breeder? First, they all give India, for the first time, nuclear weapon material — plutonium — that India says is unrestricted. As a result, they destabilize South Asia and directly threaten Pakistan. Second, it is very doubtful whether this material is really unrestricted. India’s great shortage of unsafeguarded heavy water means that India owes some explanations to its suppliers. If India refuses to place the Madras and Dhruva plutonium under safeguards, India should explain where it got the heavy water to run those reactors. Third, the plutonium bred by the Breeder will be made, at least for several years, by using plutonium from the CIRUS reactor. It thus falls squarely within the guarantee that neither CIRUS nor its heavy water will be “used” for any military purpose.
In sum, India owes more duties to its nuclear suppliers than it admits. And the duties, in the aggregate, are enough to pull its program back to peaceful use. But the duties will have to be enforced.
THE TARAPUR REACTORS
Extracting Tarapur Plutonium
Two of the thorniest problems that India has with the United States concern the Tarapur reactors. The first, which was mentioned above, is the dispute over what happens in 1993. The second is India’s claim that it can extract Tarapur’s plutonium. India says that it has the right to do so now. The United States says that India has no such right without U.S. consent. Extraction — called reprocessing — now would give India about 1,300 kilograms of plutonium, enough for 160 bombs.
Under the Tarapur agreement, India cannot extract plutonium unless the United States finds that safeguards can be “effectively applied” to the extraction plant. India says that this finding is no longer necessary, because of two facts. First, India, the International Atomic Energy Agency and the United States agreed in 1971 that the IAEA would administer U.S. safeguards rights in India. Second, in 1980 the IAEA decided that safeguards could be applied to the extraction plant and signed an agreement with India to that effect. The plant, called PREFRE (Power Reactor Fuel Reprocessing Plant), is now extracting spent fuel from the Rajasthan reactors, and the IAEA is safeguarding (inspecting) the plutonium. India contends that if the IAEA can safeguard Rajasthan plutonium at PREFRE, the IAEA can safeguard Tarapur plutonium at PREFRE.
Is India correct? Have U.S. rights been superseded? The answer is no. According to the IAEA, safeguards are effective if they detect a diversion of nuclear material within the time needed to convert it to a bomb. Detection is “timely” if it occurs within the “conversion time” of the material being inspected. The conversion time for plutonium oxide, the material coming out of PREFRE, is 1 to 3 weeks.
Timely detection is not the same as timely warning. The IAEA will not warn anyone of a diversion within the conversion time; it will only detect a diversion within that time. When the IAEA finds material unaccounted for, there is a period of time for technical evaluation, a period to report it to the Director General, and a period for him to report it to the Board of Governors, who then meet and decide: (1) whether to ask the country concerned to remedy the discrepancy; (2) whether to report the discrepancy to all members of the IAEA; or (3) whether to report it to the United Nations. This process will obviously take longer than 1 to 3 weeks. It has been estimated to take 6 months.
Timely warning is different. Its purpose is to prevent bomb making. For warning to be timely, a diversion must be detected early enough to allow other countries to intervene in time to stop manufacture. Under U.S. law, the United States, before allowing foreign extraction of U.S.-origin plutonium, must consider whether the United States will receive “timely warning . . . of any diversion well in advance of the time at which the. . . state could transform the diverted material into a nuclear explosive device.” At PREFRE, the IAEA could never warn the United States of a diversion within the 1 to 3 weeks it would take to put the plutonium into a bomb.
There is another problem at PREFRE, peculiar to plutonium extraction. The theoretical limit of accuracy in measuring the throughput of an extraction plant is plus or minus 1 percent. The practical limit can be as high as 10 percent. This means that 13 of Tarapur’s 1,300 kilograms of plutonium could be diverted without detection within the theoretical limits of accuracy. Within the practical limits, 130 could be diverted. The former amount could make 2 fission bombs; the latter, between 15 and 20.
Skepticism about inspection at PREFRE was confirmed in 1983. A news article in October disclosed that six kilograms of plutonium were missing. According to the article, uncertainties in measuring plutonium at PREFRE were running at a level of 10 percent. Moreover, after the loss was discovered, India withheld it from IAEA officials and did not acknowledge it until the LAEA inspectors demanded details. The missing plutonium was about enough for one bomb.
The last and most telling point against extraction is that India does not need Tarapur’s plutonium for civilian purposes. The Rajasthan, Madras and CIRUS reactors will supply more than enough plutonium to fuel the Fast Breeder. The Breeder and Tarapur are the only reactors in India that could use plutonium fuel; all the others use natural uranium. Thus, the only legitimate need for Tarapur’s plutonium is to make fuel for Tarapur itself. But India has no facility for making such fuel, and even if it did, it would cost far more than the low-enriched uranium that Tarapur now uses.
For all these reasons, the United States can find that safeguards would not be effective at Tarapur. Extraction would prevent timely warning, remove the last time barrier between spent reactor fuel and weapons, and create a large plutonium stockpile having no peaceful purpose.
Perpetuity: Are There Rights After 1993?
Safeguards in “perpetuity” mean that inspection and the pledge of peaceful use continue as long as the recipient has the export. India says that all rights end in 1993, when the Tarapur agreement expires. The State Department says that perpetuity is inherent in peaceful cooperation, and that the agreement legally imposes it.
The agreement provides in Article X that it “shall remain in force for a period of thirty (30) years.” One can read this to mean that not a single provision remains in force a day longer. Safeguards, peaceful use and the U.S. right to approve extraction would all end in 1993. But one can also read Article X as only limiting the obligation to sell and buy reactor fuel. Since the agreement is silent, the issue becomes one of interpretation. Everyone agrees that, in interpreting agreements, one seeks the intention of the parties.
Did the parties intend India’s meaning? If they did, the agreement would work as follows:
(1) In 1993 the Tarapur reactors could be converted to military production.
(2) In 1993 India could use all of Tarapur’s plutonium for bombs, which would mean that the Tarapur reactors would have been military production reactors for their entire lives.
(3) After 1993, India could transfer the Tarapur plutonium, or weapons made from it, to any country or group India chose.
(4) Enriched uranium shipped during the early years of the agreement would carry safeguards for almost 30 years, but enriched uranium shipped during the later years would carry safeguards for almost no time at all. Enriched uranium shipped in 1987, loaded in the reactor in 1989, discharged in 1992 and cooled for a year in a storage pond would have formed plutonium to which no safeguards would ever apply.
(5) If India stored Tarapur’s spent fuel until 1993, the U.S. right to approve plutonium extraction would never arise.
The parties could not have intended these results. The United States does not supply fuel or reactors for military production. India could not have thought the United States was doing so. Nor does the United States allow the unrestricted retransfer of U.S.-origin plutonium to other countries or groups. India could not have believed the United States was doing that either. Under India’s interpretation, the United States would have cared about peaceful use and inspection only of fuel shipped during the first part of the supply period. Toward the end, the United States would progressively lose interest. From 1986 onward, according to India, the United States would have intended to ship uranium with the knowledge that plutonium made from it would become unrestricted as soon as it left the reactor storage pool. And finally, the United States would not care about plutonium extraction unless it happened before 1993. This view is untenable, and obviously destroys rights that the agreement meant to confer.
For the Tarapur agreement to make sense, the 30-year clause must be squared with the rest of it. The only way to do that is to interpret the clause as a limit on the fuel supply commitment. Thirty years is the expected life of the Tarapur reactors. Since the reactors can be operated only with U.S.-origin fuel, it is logical for the fuel commitment to equal the reactors’ expected lifetime. By interpreting Article X this way, one avoids defeating other rights that obviously must survive it. If the 30-year provision ended all rights, the agreement would be an option contract for weapons. The recipient country would earn the right to weapons simply by performing to the end. The parties cannot have intended that result, and since they did not, India’s position is clearly wrong.
The French Takeover
France took over the fuel supply to Tarapur in 1982. After 1980, the Nuclear Non-Proliferation Act (NNPA) banned the further export of U.S. reactor fuel to a non-nuclear weapon state that had not put all of its facilities under international inspection — inspection known as full-scope safeguards. India refused. India warned the United States that a halt in the fuel supply would breach the agreement, cancel safeguards and end peaceful use. India was probably bluffing. Its legal position was weak and it would have faced even stricter controls from other suppliers. But the United States got France to take over the fuel supply, and a showdown was avoided.
Under the takeover, the United States and India waived their rights to supply and receive Tarapur’s fuel, and France agreed to complete the supply schedule. These were the only changes. France’s supply was “within the framework” of the Tarapur agreement, and peaceful use, safeguards and plutonium extraction controls remained, as well as India’s promise, made in 1974, to restrict Tarapur’s plutonium “exclusively to the needs . . . of the Tarapur Station.” India received the first French shipment in May 1983.
There are two ways to view this change. Either the United States has dropped out of the picture and France has a new contract with India, or the United States remains in the contract and has only delegated the duty to supply fuel.
The latter has clearly happened. Under the documents exchanged, the United States still controls everything required by the agreement. It controls the right to approve plutonium extraction, to maintain safeguards (through the IAEA) and to insist upon peaceful use. Moreover, if France should halt deliveries, India would look to the United States for performance. India has not released any of the U.S. duties. Thus, there is only a delegation of duty, not a new agreement. India has agreed that the United States may delegate the duty of fuel supply to France; France has agreed to assume that duty; and that is all. France has not received any of the U.S. rights, nor has France promised India anything more than the United States promised.
This point has important implications. It means that if India breaches the agreement, the United States can suspend or terminate France’s performance. Extracting Tarapur plutonium without consent would be such a breach, and so would failure to maintain safeguards or to keep the pledge of peaceful use. It would also be a breach for India to repudiate future duties.
The French takeover has therefore changed very little. In a sense, this result was required by the position of France. The Nuclear Suppliers’ Guidelines, to which France adheres, require safeguards in perpetuity for the export of reactor fuel to a non-nuclear weapon state. If France’s supply to India had been a new commitment, it would have violated France’s pledge to abide by the Suppliers’ Guidelines unless France declared that perpetuity was inherent in the U.S.-India agreement. France avoided this problem by agreeing only to serve as delegee. Consequently, the United States still controls the agreement, including France’s performance.
The French takeover also raises questions under U.S. law. If France did not make a new commitment, and France’s performance is still the performance of the United States, the United States may be violating the Nuclear Non-Proliferation Act. Either the United States has dropped out of the agreement and is complying with the NNPA because the French have taken over the contract, or the United States has not dropped out and is circumventing the NNPA through supply by a delegee. The present position of the United States is that it retains all rights under the agreement — the French and the Indians agree with this — but is not really “exporting” the fuel. The NNPA requires full-scope safeguards “as a condition of continued United States export of . . . special nuclear material.” Is a shipment by a foreign delegee of the United States from a point outside the United States an “export” if done to perform a U.S. agreement for cooperation? Is it “cooperation”? Does the NNPA intend to end “cooperation” as well as “exports”? Since exports of enriched uranium cannot be made except through cooperation, it certainly seems so. Indeed, the State Department’s principal lawyer on these matters reads it this way; so does the NNPA itself in section 405, which allows “cooperation” to continue during a grace period; and so does the House report on the NNPA.
The United States is still “cooperating” with India. This is clear. The United States stands as guarantor of France’s fuel deliveries; the United States insists that all its rights under the agreement remain in effect; the United States bases these rights upon France’s performance as delegee; and the United States has arranged for France to follow U.S. instructions on plutonium extraction, as well as other U.S. interpretations of the agreement. India’s duties run to the United States, not to France. Deliveries by France have not and could not take place without U.S. consent. In effect, the United States is continuing its agreement with India through a delegee. The State Department has said specifically that it does “not view the U.S.-India exchange as amending the U.S.-India Agreement for Cooperation or creating a ‘new’ agreement.” Therefore, by cooperating with India, the United States seems to be violating the NNPA.
REMEDIES IN THE NUCLEAR TRADE
What can the United States and India’s other suppliers do to enforce their rights? There are two choices. They can either pursue the remedies in the nuclear trade — which are limited, and depend upon what India does — or decide that nuclear arms in South Asia are such a threat that all diplomatic levers will be used to stop them.
Acts Triggering Remedies
Three acts by India could trigger remedies: India could extract Tarapur plutonium without consent, set off another test or deploy a nuclear weapon. The first would breach the Tarapur agreement. Could India be brought before the International Court of Justice? The answer is no. The jurisdiction of the Court is by consent and there is no reason why India would consent.
Could the United States take action on the Tarapur loan? The U.S. Agency for International Development (AID) lent India $ 80 million in 1963 to build the Tarapur reactors. India still owes $ 45.8 million. A breach of the Tarapur agreement entitles the United States to immediate repayment. If India does not pay, it defaults on all other AID loans. There is a standard clause under which failure to make a required payment on one AID loan is a default under all others to that country. AID can suspend disbursements under current loans and accelerate the principal due on all past loans. Suspending disbursements would cut off about $ 248 million in U.S. foreign aid to India. Acceleration would affect about $ 2.4 billion of principal still outstanding. India could avoid the effects on other loans by paying the $ 45.8 million due on Tarapur.
How effective would this remedy be? First, $ 248 million in current disbursements is puny compared to India’s total foreign aid. India now gets about $ 2 billion per year from all sources. The $ 2.4 billion in outstanding debt is more significant. But accelerating this amount would strain bilateral relations to the limit. India would say that it had not breached the Tarapur agreement, that the Tarapur loan was not due and that the $ 2.4 billion was not due either. The United States could protest but would have to settle for its unilateral remedy of suspending disbursements. Even this would carry some risk, because the cash coming into the United States from India’s loan repayments now exceeds the cash going out to India in new disbursements. Other remedies, discussed below, would be more effective.
If India deployed a nuclear weapon, it would probably not say which plutonium it used. If the bomb were made with Rajasthan or Tarapur plutonium, IAEA safeguards and India’s pledges to just about everyone would be breached. That probably will not happen. If India used CIRUS plutonium, it would breach the peaceful use pledge to Canada for the reactor and to the United States for the heavy water. Because CIRUS plutonium is not being inspected, however, there is no way to tell the difference between it and Madras or Dhruva plutonium, which, according to India, is unrestricted.
Regardless of the plutonium used, deploying a weapon would fundamentally change U.S.-India nuclear trade relations. Under international law, it would allow the United States to suspend or terminate the Tarapur agreement. When circumstances change radically from those assumed as the basis for agreement, and make one party’s performance fundamentally different, then performance is excused unless the party assumed the risk that the change might occur. This is the doctrine of rebus sic stantibus, or “fundamental change of circumstances,” contained in Article 62 of the Vienna Convention on the Law of Treaties.
It is fairly clear that if India deployed a weapon, it would be a “fundamental change of circumstances.” India would have decided to make bombs from its civilian nuclear program. After a change of this magnitude, the essential purpose of the United States — to aid India in the peaceful use of nuclear energy — could no longer be achieved. The United States could therefore suspend its performance under Article 62.
If India set off another nuclear test, India would say the device was peaceful. It would probably not say which plutonium was used. Nevertheless, another test would be a “fundamental change of circumstances” under the Vienna Convention. A test is indistinguishable from a bomb. It shows that the country has taken the road to weapons, and cannot be squared with a peaceful nuclear import program. Like deployment, it would allow the United States to suspend the Tarapur agreement and stop the fuel supply. U.S. domestic law also forbids nuclear “exports” to a non-nuclear weapon state that detonates a nuclear explosive device, “peaceful” or not. If ending exports means ending cooperation, as argued above, then U.S. law would require the United States to stop the French fuel deliveries.
Perpetuity: The Heart of the Matter
India says all rights end in 1993. What this means has already been described. The Tarapur reactors would have spent their lives making unrestricted plutonium. That plutonium could be extracted without U.S. consent and transferred to anyone, or made into weapons that could be transferred to anyone. No records would be kept; no one would know when the plutonium was separated, how it was stored, whether it had been stolen or anything else about it. The United States would have set a terrible precedent.
What can the United States do? By rejecting perpetuity, India has already repudiated an obligation that it assumed under a fair reading of the Tarapur agreement. In effect, India has declared that it will not perform when the time comes. That is a present breach under established contract law. Each party to a contract is entitled to assurance that the other will perform. When a party has reasonable grounds to believe that the other may not perform, it can demand assurances, and if they are not received, suspend performance. The United States has the right to three assurances from India: that India will never use the Tarapur reactors or their fuel for any military purpose; that India will never extract Tarapur plutonium without U.S. consent; and that India will maintain safeguards on the Tarapur reactors and their fuel as long as either is in India. India has refused repeatedly to give these assurances — which is a clear repudiation. It therefore allows the United States to suspend performance under international law.
What are the risks of this remedy? India could stick to its rejection of perpetuity, which would deny that the United States was entitled to assurances or to suspend performance, and cause India to claim that the United States itself had breached. India could say that after such a breach, it was entitled to extract the plutonium, end safeguards and forget about peaceful use.
How would the United States respond? It could offer to go before the International Court. But if India refused, the parties would be at loggerheads. Each would deny the rights of the other, and India would have to give in or face an end to the fuel supply. Would France end it under such conditions? France has agreed to follow U.S. interpretations of the Tarapur agreement. It is bound to as delegee. If France continued deliveries after being asked to stop, it would amount to a new supply commitment between France and India. But under its adherence to the Nuclear Suppliers’ Guidelines, France cannot make such a commitment without requiring perpetuity. Neither can any other potential supplier; they all adhere to the Guidelines. The Guidelines require that, in the event of an “illegal termination or violation of IAEA safeguards by a recipient, suppliers should consult promptly . . . [and] pending the early outcome of such consultations . . . not act in a manner that could prejudice any measure . . . adopted by other suppliers concerning . . . that recipient.” Since neither France nor another supplier would be likely to “prejudice any measure” adopted by the United States, France would follow its delegor’s order and other suppliers would stay out of the fight. India’s choice would be limited: provide assurances or close Tarapur.
Should the United States force such a choice? If the Indians are truthful about what they will do in 1993, what can the United States look forward to? Six years from now, the Indians will do what they wish with U.S.-supplied plutonium. Why continue to supply until then? So India will have more? No one knows whether India will ever want Tarapur’s plutonium for bombs; even India may not know. Nonetheless, India will still have the plutonium. If the United States allows its delegee to continue deliveries, there will simply be more of it. Would the Indians close Tarapur rather than give in? Perhaps. If they did, they would probably extract and stockpile the plutonium. But that is what they say they will do anyway after 1993. Does it make a difference whether it happens now or then?
The present U.S. policy is to continue to supply and hope India will come around to the U.S. view by 1993. But what strategy does the United States have if India does not? The worst possible result would be for India to get deliveries up to 1993 and then repudiate: India will have got its way costfree, and shown the world that the United States cannot enforce its agreements. The United States must ask itself what leverage it will have when the last shipment arrives. As long as the supply continues, India can simply wait. Consequently, there appear to be only two alternatives: a showdown now or a drift to 1993. Would the showdown be better?
With a showdown, there is at least a chance that India will give in. India may find electricity from Tarapur, and good relations with the United States, more valuable than its strained reading of the agreement. When India could not operate the Rajasthan reactors without imports, it accepted safeguards in perpetuity to get Soviet heavy water. If, however, India prefers to close Tarapur now rather than accept perpetuity, how likely will India be to accept perpetuity in 1993, when the reactors will be closing anyway? As 1993 approaches, the reactors approach the end of their useful lives, and India’s cost becomes smaller and smaller. Time erodes the U.S. position.
For all these reasons, the nuclear trade remedies are limited. If a country misuses a nuclear export, one cannot do much more than halt further nuclear exports. One may have the “right” to get back what one supplied — the Tarapur agreement allows the United States to get back what it supplied if India breaches — but one cannot dismantle and retrieve a reactor. Or even, in the case of Tarapur, get back the spent fuel. There is a shortage of available shipping casks, and limits on the railroad from Tarapur to Bombay. It would take many years and great expense to bring Tarapur’s spent fuel back to the United States. Nor can one retrieve technology. The United States has trained more than one thousand Indian scientists; there is no way to make them forget what they have learned.
GENERAL DIPLOMATIC REMEDIES
Diplomatic remedies can touch every aspect of India’s relations with the United States. And they can touch India’s relations with other countries and international institutions. In 1984-1985, India was allowed to buy between $ 1 and $ 2 billion per year in high-technology U.S. products — two or three times as much as it was allowed to buy in 1983-1984. These purchases include the technology for building computers and fighter plane engines. The United States is also considering India for a supercomputer, whose export is sensitive militarily and a mark of high favor. All this could be stopped.
If India and Pakistan continue to do what they are doing now, the following will happen: Pakistan will be able to produce enough high-enriched uranium within the next few years for several bombs. It may perform a test, or assemble weapons quietly. It seems to have the competence to do either. Pakistan will have done this for fear of India, which, meanwhile, will have declared as unrestricted — and therefore free for weapons — the plutonium coming out of three heavy-water reactors and one fast breeder test reactor. The plutonium will be enough for a small arsenal, and will have been made either with secret imports, with heavy water diverted from safeguards, or by using the restricted CIRUS plutonium as fuel. At Tarapur, India will continue to build up a large plutonium stockpile, which it says will be free for bombs in 6 years. This is what will happen if nothing else happens. What are the implications?
First, the next Indo-Pakistani war may be fought by enemies with nuclear weapons. If it is, the world will be forced to learn whether a conventional war between nuclear-armed rivals is possible. It may not be. If it is not, and the superpowers are dragged in, things could go badly. Second, the effort to control proliferation will have been flouted by two important countries. If it cannot survive this test and begins to unravel, one can expect regional nuclear confrontations in several places. All this is worth some trouble to avoid. It still can be avoided — just barely — if the United States, Canada and the Soviet Union act decisively.
These countries can do a number of things. First, they can insist upon the broadest reasonable interpretation of India’s duties under the nuclear trade agreements. This means that India owes the United States a pledge of peaceful use and safeguards in perpetuity for Tarapur plutonium. It means that India owes the United States and Canada the peaceful use of CIRUS plutonium and any plutonium the Breeder makes from CIRUS plutonium. It means that India owes the United States, Canada and the Soviet Union a duty to safeguard and restrict to peaceful use the plutonium from Madras and Dhruva unless India can prove it made that plutonium with unsafeguarded heavy water. If India were to perform these duties, its nuclear program would resume a peaceful course. One could then hope for similar behavior from Pakistan.
If India rejects the duties, the nuclear trade remedies could be pursued. The United States could cut off Tarapur’s fuel and make the Tarapur reactors worthless to India for their remaining lives. India cannot make the fuel for Tarapur and no other supplier would sell to India after a U.S. cutoff. Canada and the Soviet Union could get their heavy water back provided that India had misused it, which Canada and the Soviets could assume to be the case if India could not explain where it got the unsafeguarded heavy water to run Madras and Dhruva. The return of Soviet and Canadian heavy water would shut down most of the reactors in India and make a shambles of its nuclear energy program. Even a halt in exports of Soviet heavy water would close some of India’s reactors. India would still have the plutonium from Tarapur, Rajasthan and CIRUS, and a small amount from Madras, but at a staggering cost. It would make the Indian bomb an astronomically expensive substitute for the Indian nuclear energy program, rather than a cheap byproduct of it.
If these costs were not enough, one could halt analogous trade. India would not get its computer technology, or its fighter plane engine technology or the supercomputer. To counter the old argument that India would be pushed into the arms of the Soviets, the Soviets would have to cooperate. They would have to cut back or stop conventional arms sales, or other transfers beneficial to India. The final step would be economic. The United States could use its influence over international lending and trade. The United States is India’s largest trading partner. The resulting costs could well be high enough to convince India that it would be weaker, rather than stronger, with the atomic bomb.
What are the chances that this will be done? They are not very great. To take effective action, the United States, Canada and the Soviet Union would have to change their way of thinking about nuclear arms proliferation. The old way — accommodation in the short term and finger crossing for the long term — has produced a nuclear arms threshold in South Asia. If the threshold is crossed, it will be because of decisions in Washington, Ottawa and Moscow.
 Woodward, Pakistan Reported Near Atoms Arms Production, Wash. Post, Nov. 4, 1986, at A1; see also L. SPECTOR, THE NEW NUCLEAR NATIONS 115-16 (1985).
 John Scali, interviewed on Good Morning America (ABC television broadcast, July 11, 1985), discussed in L, SPECTOR, supra note 1, at 120-21.
 Interview in Le Monde, June 4, 1985.
 See 22 U.S.C. § 2429a (1982); see also International Security and Development Cooperation Act of 1985, Pub. L. No. 99-83, 99 Stat. 190 (primarily codified in various sections of 22 U.S.C. (Supp. III 1985)).
 Agreement for Cooperation Concerning Civil Uses of Atomic Energy, Aug. 8, 1963, United States-India, 14 UST 1484, TIAS No. 5446, 488 UNTS 21 [hereinafter Tarapur agreement].
 Assuming 8 kilograms per bomb. In normal operation, Tarapur produces “reactor grade” plutonium, which contains about 20% Pu 240, an isotope that can lower the yield of a weapon. However, the open literature on bomb design assumes such weapons are reliable. A fledgling nuclear weapon state may be content with lower yield weapons.
 “Weapon grade” plutonium is formed only in fuel taken out of the reactor after a short irradiation time. It is called “low burnup” fuel. At the burnup that was average up to 1977, Tarapur’s plutonium would be 85-90% free of undesirable isotopes. M. BENEDICT, T. PIGFORD & H. LEVI, NUCLEAR CHEMICAL ENGINEERING 88 (2d ed. 1981). Also, there is a variation in burnup across the reactor’s core. Id. at 94. Finally, the burnup varies along the length of each fuel rod. Id. at 111. It is weapon grade on the ends. Id. According to newspaper accounts, India could cut off the low-burnup ends and extract that plutonium separately. India Makes Another Bomb, Sunday Observer (Bombay), Aug. 30, 1981. The result is that Tarapur’s spent fuel contains a fair amount of weapon grade, or near-weapon grade plutonium. By grouping the low burnup rods for separate reprocessing, one could achieve a degree of purity of about 90%. By cutting off the ends of the rods, one could achieve a higher purity than that. If only 10% of the 1,800 kilograms of plutonium produced by 1993 were lightly irradiated, that would make 180 kilograms of weapon grade, or near-weapon grade plutonium — enough for 36 bombs at 5 kilograms per bomb. Hildenbrand, Fast Critical Masses of Fissile Material for Nuclear Explosives, in NUCLEAR PROLIFERATION FACTBOOK 295 (1985) (paper presented at Atomic Industrial Forum Conference on International Commerce and Safeguards for Civil Nuclear Power, March 1977).
 For histories of the Indian nuclear program, see generally R. WOHLSTETTER, THE BUDDHA SMILES: ABSENT-MINDED PEACEFUL AID AND THE INDIAN BOMB (1977); S. BHATIA, INDIA’S NUCLEAR BOMB (1979); AND D. HART, NUCLEAR POWER IN INDIA (1983).
 Letter from Myron B. Kratzer, United States Department of State, to Benjamin Huberman, United States Nuclear Regulatory Commission (July 20, 1976).
 NUCLEONICS WEEK, July 1, 1976, at 6-7; R. WOHLSTETTER, supra note 8, at 153-54.
 India has no right to “substitute” other heavy water for the U.S. water in CIRUS through a bookkeeping transaction. Many nuclear trade agreements give such rights, which allow the recipient to consider water supplied under restrictions as being in whichever facility is convenient, as long as safeguarded heavy water is in a safeguarded facility. No such right was given to India. Agreement, Mar. 16, 1956, United States Atomic Energy Commission-India, § 9.
 The Canadian agreement is described in Hunt, Canadian Policy and the Export of Nuclear Energy, 2 U. TORONTO L.J. 69, 77 (1977).
 Milhollin, Dateline New Delhi: India’s Nuclear Cover-up, 64 FOREIGN POL’Y 161 (1986).
 Agreement of Jan. 27, 1971, United States-India-International Atomic Energy Agency, secs. 4, 12, 22 UST 200, TIAS No. 7049, 798 UNTS 115.
 INTERNATIONAL ATOMIC ENERGY AGENCY, THE STRUCTURE AND CONTENT OF AGREEMENTS BETWEEN THE AGENCY AND STATES REQUIRED IN CONNECTION WITH THE TREATY ON THE NON-PROLIFERATION OF NUCLEAR WEAPONS, IAEA DOC. INFCIRC/153 (Corrected) (Vienna 1972). India is not a member of the Treaty (done July 1, 1968, 21 UST 483, TIAS No. 6839, 729 UNTS 161) and so is covered by IAEA Doc. INFCIRC/66/Rev.2 (Sept. 16, 1968). The IAEA takes the position, however, that its current practice is to apply INFCIRC/153 even to non-NPT countries. INTERNATIONAL ATOMIC ENERGY AGENCY, IAEA SAFEGUARDS, AN INTRODUCTION (Vienna 1981). See also Herron, A Lawyer’s View of Safeguards and Non-Proliferation, 24 IAEA BULL. No. 3, 1982, at 34-35.
 INTERNATIONAL ATOMIC ENERGY AGENCY, IAEA CONTRIBUTION TO INFEC 7 (Vienna 1979).
 Herron, supra note 17, at 37.
 A. wohlstetter, Addendum H to R. WOHLSTETTER, supra note 8, at 236. A. Wohlstetter provides a thorough discussion of the consequences of allowing reprocessing of Tarapur plutonium.
 Nuclear Non-Proliferation Act of 1978, Pub. L. No. 95-242, § 303(a), 92 Stat. 120, 127-31 (codified as amended at 42 U.S.C. § 2160 (1982)) [hereinafter Nuclear Non-Proliferation Act].
 INTERNATIONAL ATOMIC ENERGY AGENCY, IAEA SAFEGUARDS TECHNICAL MANUAL 36 (Vienna 1976).
 See e. g., U.S. GENERAL ACCOUNTING OFFICE, REPORT NO. EMD-80-38, NUCLEAR FUEL REPROCESSING AND THE PROBLEM OF SAFEGUARDING AGAINST THE SPREAD OF NUCLEAR WEAPONS 13 (1980).
 Sunday Observer (Bombay), Oct. 16-23, 1983, at 1.
 Nuclear Non-Proliferation Act, supra note 21, § 306 (codified at 42 U.S.C. § 2157 (1982)).
 For a discussion of the Tarapur problem and its history, see Clausen, Nonproliferation Illusions: Tarapur in Retrospect, ORBIS, Fall 1983, at 741.
 Cable New Delhi 22, 789 (Nov. 29, 1982); Note to the United States from India (Nov. 30, 1982); Note to India from the United States (Nov. 30, 1982) (on file at U.S. Dep’t of State).
 Times of India (Bombay), May 7, 1983, at 1.
 Personal communication with Bertrand Barre, Nuclear Attache, Embassy of France (Feb. 2, 1984). In the negotiations on the takeover, the French tried to get India to agree specifically that pursuit and perpetuity would attach to fuel deliveries, but the Indians resisted and the question was left open. W. DONNELLY & N. MILLER, TERMINATION OF U.S. NUCLEAR COOPERATION WITH INDIA (Cong. Research Serv. Issue Brief No. 81,087, 1983); NUCLEONICS WEEK, Dec. 2, 1982, at 1.
 This is elementary contract law. See, e.g., RESTATEMENT (SECOND) OF CONTRACTS § 318 (1979). It applies to international agreements. Article 60 of the Vienna Convention, note 47 infra, provides that “[a] material breach of a bilateral treaty by one of the parties entitles the other to invoke the breach as a ground for terminating the treaty or suspending its operation in whole or in part.” Article 2(1) defines a treaty as “an international agreement concluded between States in written form . . . whatever its particular designation.”
 INTERNATIONAL ATOMIC ENERGY AGENCY, INFCIRC/254 APPENDIX: GUIDELINES FOR NUCLEAR TRANSFERS § 4 (1978), incorporating the durational requirements for safeguards stated in IAEA Doc. GOV/1621 (1973).
 Nuclear Non-Proliferation Act, supra note 21, § 306 (codified at 42 U.S.C. § 2157 (1982)).
 Atomic Energy Act of 1954, ch. 1073, § 123, 68 Stat. 930, 940 (1954) (codified as amended at 42 U.S.C. § 2153 (1982)).
 Bettauer, The Nuclear Non-Proliferation Act of 1978, 10 LAW & POL’Y INT’L BUS. 1105, 1125 (1978), stating: “the full scope safeguards requirement mandates termination after a stated time limit of U.S. cooperation with those recipients that do not meet the requirement.”
 42 U.S.C. § 2153d (1982).
 H.R. REP. NO. 587, 95th Cong., 1st Sess. 12 (1977).
 Letter from Powell A. Moore, U.S. Dep’t of State, to Richard L. Ottinger, U.S. House of Representatives (Jan. 27, 1983).
 Parties may consent by a special agreement referring a dispute to the Court, by a treaty or convention preceding the dispute that refers such disputes to the Court or by a declaration accepting compulsory jurisdiction over the dispute under the “optional clause” of Article 36 of the Court’s Statute. Statute of the International Court of Justice, 59 Stat. 1055 (1945), TS No. 993, Art. 36; 2 L. OPPENHEIM, INTERNATIONAL LAW: A TREATISE 58-59 (H. Lauterpacht 7th ed. 1952). None of these conditions is met here.
 OFFICE OF FINANCIAL MANAGEMENT, AGENCY FOR INTERNATIONAL DEVELOPMENT, STATUS OF LOAN AGREEMENTS 70 (1985).
 Loan Agreement Between the President of India and the United States of America, Dec. 7, 1963, AID Loan No. 386-H-091. The loan is discussed in R. WOHLSTETTER, supra note 8, at 82; and in D. HART, supra note 8, at 39.
 AGENCY FOR INTERNATIONAL DEVELOPMENT, STANDARD FORM LOAN PROVISIONS ANNEX § D.3(d).
 Id. § D.2(c).
 OFFICE OF FINANCIAL MANAGEMENT, supra note 40, at 78.
 AGENCY FOR INTERNATIONAL DEVELOPMENT, CONGRESSIONAL PRESENTATION FOR FISCAL YEAR 1984, Ann. II, Asia, at 43.
 Vienna Convention on the Law of Treaties, opened for signature May 23, 1969, UNTS Regis. No. 18,232, UN Doc. A/CONF. 39/27 (1969), reprinted in 63 AJIL 875 (1969), 8 ILM 679 (1969).
 22 U.S.C. § 2429a (1982).
 See, e.g., RESTATEMENT (SECOND) OF CONTRACTS § 250 (1979); U.C.C. § 2-610 (1978).
 RESTATEMENT (SECOND) OF CONTRACTS § 251 (1979); U.C.C. § 2-609 (1978).
 India has argued that safeguards apply to the Tarapur reactors only because they use fuel supplied under the U.S. agreement. This is clearly wrong. The operative language of Article VI of the agreement gives the United States the right to review the design of any facility used to separate plutonium “produced in the Tarapur Atomic Power Station.” This catches plutonium furnished by non-U.S. suppliers, and therefore means that the reactors themselves are safeguarded. Article VI, para. B(2), which requires a system of records, is written the same way. It catches all material, from whatever supplier, “produced in . . . Tarapur.” In addition, the pledges of peaceful use in Article VIII, and the right to approve extraction in Article II, catch material produced with all U.S. exports, including the reactors.
 The Vienna Convention, supra note 47, provides in Article 60 that a “material breach of a bilateral treaty by one of the parties entitles the other to invoke the breach as a ground for terminating the treaty or suspending its operation in whole or in part.” It also provides that a “material breach . . . consists in . . . a repudiation.”
 IAEA Doc. GOV/1621, supra note 32.
 Id. § 14.
 Wash. Post, Feb. 7, 1986, at A21.
 N.Y. Times, Oct. 1, 1986, at A15; Wash. Post, Jan. 7, 1987, at A16.
 Wash. Post, July 8, 1986, at D1; and Dec. 12, 1986, at A45.
Gary Milhollin is Professor of Law, University of Wisconsin Law School.