POST-START: WHAT DO WE WANT? WHAT CAN WE ACHIEVE? by Richard L. Garwin IBM Research Division Thomas J. Watson Research Center P.O. Box 218 Yorktown Heights, NY 10598 (914) 945-2555 (also Adjunct Professor of Physics, Columbia University; Adjunct Research Fellow, CENTER FOR SCIENCE AND INTERNATIONAL AFFAIRS Kennedy School of Government Harvard University) February 27, 1992 TESTIMONY to the Committee on Foreign Affairs United States Senate (Published in Hearings Before the Committee on Foreign Relations, One hundred Second Congress, Second Session), pp. 130-147). ABSTRACT START should be ratified and implemented as soon as possi- ble. The post-START era is well under way with the initi- atives set in motion by President Bush September 27, 1991, and disablement and guarded storage of many thousands of nu- clear warheads must be our urgent near-term objective. Post-START forces should have an early goal of 3000 actual strategic warheads on each side, plus perhaps 400 air- delivered tactical weapons. As soon as other nuclear powers see merit in reductions, the U.S. and Commonwealth should move each to 1000 single-warhead weapons. The threat of ac- cidental or unauthorized launch should be handled by instal- lation of a passive destruct-after-launch system rather than by active defenses, and theater-range missiles should be eliminated by extending to all nations the INF Treaty which binds the U.S. and the successor states of the Soviet Union. (Embargoed until 1000, Thursday February 27, 1992.) R051TEST 022092TEST 02/24/92 Views of the author, not of his organizations INTRODUCTION I am pleased to have the opportunity to testify on Post-START. Two Committee reports of last spring and last fall, "START: Present Status and Prospects," and "The START Treaty in a Changed World" provide an excellent background, so that I will not repeat much of the factual material and analysis shown there. It is almost 20 years since I testi- fied in support of the ABM Treaty and the SALT Limited Of- fensive Agreement. Had we been steadfast at that time in our purpose and in our understanding of the perils posed to the world by large numbers of Soviet nuclear warheads, we could well have avoided the tremendous buildup of nuclear weaponry on the two sides, at great savings in treasure and years of exposure to that hazard. Had we had the imagina- tion and guts ten years ago that have been evident in the last year, START could have been fully implemented by now, rather than just beginning. In any case, it is important to put into force the START Treaty, not only to achieve the modest benefits of the re- ductions for which it provides, but especially to have the detailed vehicle of obligation and mechanisms which will en- able and support further reductions and limitations which are so much in the interest of U.S. and world security. In his testimony last fall, Dr. Panofsky ably presented the substance of the report of the National Academy of Sciences Committee on International Security and Arms Control (CISAC) of which I was also an author. That unanimous report lays the foundation for moving beyond START, and was soon fol- lowed by the initiative of President George Bush and that of President Mikhail Gorbachev toward the unilateral reduction, elimination, and in many cases the destruction of nuclear weapons-- especially tactical nuclear weapons which would have been untouched by START itself. Those initiatives were wise but long overdue, since it is not the intended purpose of nuclear weapons (to fight on the battlefield, for in- stance) which poses the greatest peril, but the fact that a nuclear weapon can destroy a city. The possession of more nuclear weapons on our part has not in any way compensated for the increased hazard posed by the enormous increase in numbers of nuclear weapons of the former Soviet Union, which, themselves, are now seen by their holders not as an asset but as a costly and dangerous liability. So a lot of attention is being given to the implementation of the Bush-Gorbachev-Yeltsin initiatives. Although this is not the primary topic of my presentation, the three items are linked-- START implementation, destruction of many thou- sands of tactical nuclear weapons, and the nature of the post-START world. It is important that the process of removing all non- strategic weapons to the Russian Republic be completed, as scheduled, by July 1, 1992, and that this process be carried out safely. Weapons to be moved are temporarily disabled, as I understand it, transported, disabled more permanently in many cases on their arrival, and stored for destruction. According to information from the other side, capabilities for the disassembly and destruction of nuclear weapons can provide routinely for 1500 such per year, but also could handle some 4500 per year, except for a shortage of adequate storage facilities for the plutonium that would be removed from the weapons. As has been indicated in recent testimony(1) by Administration witnesses, there is active discussion between the U.S. and the Commonwealth of Inde- pendent States to identify adequate storage capabilities, secure transport capabilities, and the like. In view of our proven inability to imagine the political future, it would be prudent to implement as quickly as possible these meas- ures that have been sketched by the Presidents, to eliminate physically weapons which are now regarded as a liability and a hazard not only to the other nations of the world but to one's own nation. Here I go beyond the CISAC report to present my own views, in the light of the political evolution thus far and of the initiative of the Presidents. I recommend that all non- strategic nuclear weapons be destroyed, except, as an in- terim measure, for fewer than 1000 (say 400) U.S. air-delivered weapons that might remain nominally at the service of NATO; an equal number of similar air-launched weapons would be retained, for the time, by the Commonwealth of Independent States, or by Russia. Previous testimony in- dicated general agreement that U.S. and Soviet strategic forces could each be reduced to the level of 3000 actual warheads, and I urge that this be done quickly. It is a welcome change to see frank discussion about what it is that bothers us, and, in some cases, even solutions that really address the problem. It was not always thus. For instance, the development of the MX 10-warhead missile (Peacekeeper) and its deployment in Minuteman silos was in no way a remedy for the silo vulnerability that was used to motivate the MX program. In his Los Angeles Times article July, 1991, former Presi- dent Reagan recalls with approval his March 1983 announce- ment of the SDI program, which, he says, called for an SDI "effective enough to stop a high percentage of incoming mis- siles from destroying our own arsenal, in order to discour- age an enemy from launching an attack in the first place." Of course, SDI had no such effect; the 1983 and 1984 Scowcroft commission reports explained why such an attack was not a matter of strategic concern; and the goals of the SDI in any case were not the (actually achievable but un- sought) goal of preserving the strategic retaliatory force but rather that of rendering Soviet nuclear weapons "impo- tent and obsolete," so that we could abandon our own nuclear weapons. As more nearly achievable goals for SDI replaced the initial purpose, it was often observed that it would be cheaper, quicker, and more reliable to eliminate half of the Soviet strategic warheads by a treaty than to attempt to nullify them in flight to their targets, and there is now consensus that such reductions are feasible, important, and urgent. SDI goals are now focused, for the present, on the achieve- ment of GPALS-- a Global Protection Against Limited Strike. This is to include protection against an unauthorized or ac- cidental launch of nuclear-armed missiles of the former Soviet Union, defense of the United States against nuclear- armed ICBMs from a third-world nation, and protection of de- ployed U.S. forces and friendly and allied cities against shorter range ballistic missiles, armed with chemical, bi- ological, or high-explosive warheads. Additional questions raised in last year's testimony include the role of nuclear testing, and specifically of space-based weaponry. I will address all of these briefly in the post-START context. REDUCTIONS IN U.S. AND COMMONWEALTH NON-STRATEGIC WARHEADS. I assume and recommend that these be transferred to secure and internationally guarded storage, in the owning nation, pending destruction, except for 400 warheads intended for air delivery. Control and verification should be aided by an exchange of declarations and continuing information, in- cluding serial numbers of weapons and their major compo- nents, with the attachment of tags of increasing reliability. Much of the benefit of such tags would be ob- tained by a simple "license plate" assigned irrevocably to a particular warhead, to aid accountability and verification. How else could one verify an assertion that, say, 9870 tac- tical warheads remain in a dozen storage sites? It is not practical to put them all on display simultaneously and to immobilize them until they are counted. Far better to use random sampling of a detailed declaration. REDUCTION OF U.S. AND FORMER SOVIET STRATEGIC WARHEADS. Weapons removed from aircraft can be restored in a couple of days or hours. ICBMs taken off alert can be back in full action in days or weeks. Those warheads slated for de- struction under START or post-START should be removed from their delivery vehicles, disabled, tagged, and transported to secure and internationally guarded storage to await their eventual destruction, as proposed by President Bush in his September 27 initiative. There is now general agreement that these reductions should be made without the necessity of building any new system. This puts the emphasis on the preferential retention of single-warhead missiles and on downloading-- the reduction of the number of warheads carried by an individual MIRVed missile. It is a limitation on warheads themselves, com- bined with individual accountability and destruction of sur- plus warheads, that permits reliable and verifiable downloading. A regime for controlling warheads, of course, must incorporate data exchange and verification, a strict control of warhead construction and remanu- facture, and must be buttressed by a cutoff of production of highly enriched uranium (HEU) and plutonium. In reality, neither the U.S. nor the Commonwealth is producing HEU, and the U.S. has produced no plutonium for several years. With the destruction imminent of many thousands of war- heads, there will be in any case no need for plutonium production in the Commonwealth, and it has been stated that it will cease within this decade; it is evidently a make-work activity. Warheads totalling a given number are no greater threat to the other side if they are housed on MIRVed missiles than on single-warhead missiles. Quite the contrary, in fact, since singlets provide better flexibility, a wider range of tac- tics, and the like. For equal numbers of warheads on the two sides, MIRVs are a problem for their owner, because a single warhead exploding near a silo housing a MIRVed missile can destroy four or six or ten warheads. It is true, however, that even a side with singlet missiles may be rendered insecure by an equal number of warheads on MIRVed missiles on the other side, but the interaction is complex: A side possessing (only) MIRVed missiles in silos will recognize the vulnerability of those missiles to accurate warheads from the other side, and, lacking effective defense, it will likely rely on Launch Under Attack (LUA) or Launch On Warning (LOW) either as a residual capability or as a de- clared strategy to deter a disarming strike. Thus, if we assume the U.S. had only singlet mis- siles in silos and the Soviets only 10-warhead SS-18s in silos (with an equal number of warheads in total) the U.S. should certainly be concerned that the Soviets might anticipate a first, disarm- ing strike from the U.S., and would be ready, therefore, to launch their missiles on warning, thereby assuredly deterring such an attack. But what happens if there is a false warning signal? The resulting critical dependence on warning systems high- lights our foolishness in introducing MIRV in the first place, and the tragedy of not proposing to ban MIRV in the SALT negotiations, when such a ban could have been verified by production and deployment limitations in the U.S. and test limitations in the Soviet Union. But in 1992, the important point is that U.S. security is not in any way impaired if Russia should retain a few MIRVed missiles, whether by agreement or clandestinely. And the new political situation permits and encourages cooperative warning systems to reduce further the instability just dis- cussed, and to limit and to verify actual numbers of nuclear warheads. Another point that is widely misunderstood involves the cost or "economics" of operating ICBM or SLBM systems with single warheads instead of their full complement of MIRVs. For in- stance, the U.S. Joint Chiefs of Staff are cited as provid- ing an illustration of a post-START initiative force of 4700 U.S. strategic warheads, of which 500 would be on Minuteman III downloaded to single warheads, 2300 on 17 Trident submarines with 432 C-4 and D-5 missiles, and 1900 on possibly 95 B-1s with 20 ALCM each. If a force of 2000 strategic warheads were to be obtained by reducing the Trident loading to 1000 total and the B-1 loading to 500 total, it is widely reported that the cost per warhead in the SLBM deployment would be excessive. In fact, the actual operating cost of 17 Trident submarines with 1000 warheads would be somewhat less than with 2300 warheads, not more. We would save money. The operating cost could be further reduced by housing the 1000 warheads on a fleet of eight in- stead of 18 submarines, but many, including myself, believe this to be "too many eggs in one basket." If one considers a strategic force for the longer term of 1000 warheads, total-- perhaps 400 on singlet ICBMs in silos, 400 on SLBMs, and 200 on ALCMs-- one would have the choice between operating 18 Trident submarines, each with 24 missiles downloaded to a single warhead, or 18 submarines, each with 20 launch tubes blocked, and the other four con- taining 6-MIRV SLBMs. The vulnerability of the SLBM to anti-submarine warfare (ASW) is identical in the two cases, unlike the situation of silo-based weapons. Nevertheless, there is a strong preference for downloading even the SLBMs to singlets, because a modern SLBM has intercontinental range, and if it remained a MIRVed missile for the long term, the infrastructure and testing of MIRVed weapons would continue indefinitely. This would constitute an unnecessary potential for breakout, involving the deployment of MIRVed SLBMs on land. So I have a strong preference ultimately for downloading the SLBMs themselves. As has been discussed in previous testimony to this Commit- tee, while many believe it would be highly desirable to re- duce the strategic warhead numbers of the former Soviet forces to 1000 for the longer term, and that it would be ac- ceptable to reduce U.S. holdings to the same number, this could not practicably be achieved without limitation and even reductions on the nuclear forces of the other three ma- jor nuclear powers, Great Britain, France, and China. I strongly believe that it would be in the interest of these three nations to commit now to begin reductions of their own holdings to a level of 300 each, when U.S. and former Soviet holdings reach 3000, so that for the longer term, with num- bers like 1000, 1000, 300, 300, and 300, each of those na- tions would have about 30% as many warheads as the U.S. or Russia, rather than one or two percent, as is now the case. A DECADES-LONG INTERVAL WITH U.S. AND COMMONWEALTH FORCES OF 1000 WARHEADS EACH? A possible future (and a desirable one compared with the 60,000 nuclear weapons we now have in the world, or alternatively with the future of massive prolifer- ation of nuclear weaponry), would be a continuation for many decades of U.S. and Commonwealth strategic forces at the level of 1000 warheads each, with no additional tactical weapons. Nuclear warheads can be maintained in working or- der by non-nuclear testing and occasional "remanufacture" to original specifications (although some would prefer another course). A nuclear weapons establishment capable of remanu- facturing, say, 200 warheads per year would be adequate for this purpose. It would need to be safeguarded against clan- destine use, to ensure that one warhead was destroyed for every one manufactured. The delivery vehicles themselves pose a very different prob- lem. They are far more complex than a nuclear warhead and, especially if downloaded to singlet configuration, are enor- mously oversized and much more costly than necessary for their future purpose. Furthermore, it makes very little sense to demand that future Soviet forces indefinitely in- clude SLBMs capable of carrying 10 warheads, even if down- loaded to one. Over the years, each side should have the option, once, to replace a large MIRVed missile by a small single-warhead missile, sized to carry the existing warhead and some suitable package of penetration aids. Eventually, not only the missiles, but also the replacement submarines could be much smaller, and perhaps more numerous, with the same total number of launch tubes in the force, each carrying a singlet missile of intercontinental range. I emphasize that there is no urgency for the development of such missiles or submarines, but that around the year 2000 a modest development program should be initiated that would result in the deployment of small missiles and smaller sub- marines around the year 2006, as a way of reducing ongoing expenditures and limiting the future launch capability (breakout potential). A COMPREHENSIVE BAN ON NUCLEAR TESTS? This question was ad- dressed by several witnesses in the 1991 hearings including Stephen J. Hadley and James R. Schlesinger. The former im- plied that we will continue testing to improve our nuclear explosives, "that is to say make them cleaner, smaller, with less radiation, and so on." On the other hand, Dr. Schlesinger states explicitly, "I do not and have not favored a comprehensive test ban. However I can see the possibility of comprehensive test limitation-- with a low thresh- old and a limited number of tests. That would serve to sustain confidence in the reliability of the stockpile-- while precluding the development of new nuclear explosive devices." I have the same goals as James Schlesinger, but having worked on nuclear weapons for many years at Los Alamos, I am confident that stockpile reliability, safety, and security can be maintained without continued nuclear testing. I would strongly advocate that the United States take the lead now in negotiating a comprehensive ban on nuclear tests, which would take full effect in 1995, and which would permit the five nuclear powers a decreasing number of small under- ground tests until that time. We need a CTBT in the fight against proliferation of nuclear weapons. This is not so much because a test ban will abso- lutely prevent a nation from developing its first nuclear weapon. This can be done, with some probability, without any nuclear test at all. But a nation violating a universal test ban puts itself outside the law or generally accepted custom. UN military action against Iraq and the scattered references in the Hearing record to "military sanctions" point the way to the ultimate response to imminent nuclear proliferation. I believe that the United States would be far more effective in leading world opinion and action against proliferators in an era in which we have universally abandoned the right to test nuclear weapons underground. As a matter of simple human nature, it is very difficult for our nuclear weapons people to believe that somehow they have a right to test nuclear weapons, while those in other na- tions do not. On the other hand, I predict that we will have very little tolerance for nuclear tests in other na- tions if we have abandoned them ourselves. GLOBAL PROTECTION AGAINST LIMITED STRIKE (GPALS)? Although it is the ends in which we are interested, the means tend to assume compelling importance, perhaps because some action is involved-- a first step toward the goal. As usual, however, it is desirable to disaggregate the goals, to see whether their elements might be better achieved (sooner, at less cost, more reliably) by a different set of means. This was the case, for instance, with the reduction of the Soviet strategic threat by negotiation rather than by active de- fense. It could have been the case in the reduction of silo vulnerability by a close-in non-nuclear defense rather than by space-based weaponry. And it is the case with the laudable aims of GPALS. Dr. Schlesinger's testimony regarding ICBMs stated "There is no third world threat and it is not going to emerge in this century;" Harold Brown had testified to the same conclusion ("aside from China"). The one-site deployment proposed by the Missile Defense Act of 1991 (for deployment by 1996) would thus have the sole purpose of defending against unau- thorized or accidental launch of former Soviet strategic weapons. This problem can be addressed far more effectively and sooner by cooperative efforts with the Commonwealth leadership to verify and to improve the integrity of the permissive action links apparently used with all Soviet strategic weapons.(2) A totally separate system that could be implemented within a year or two and would assuredly find a willing partner on the other side would be an enhancement to the PAL, with might actually be separate and could log- ically be called a passive Destruct After Launch (p-DAL).(3) This would capitalize on a system that has traditionally been available on deployed Soviet operational missiles, ca- pable of destroying the missile after launch in case it is not responsive to its guidance system. The p-DAL would un- conditionally energize the destruct mechanism a few seconds after launch, unless an appropriate coded signal had been received by the missile just before launch from its silo or launch tube. Buried inside the missile, such a system could not readily be disconnected by the operating crews, and it could be arranged so that an attempt to disconnect it would irrevocably disable the missile (as is the case with the more modern U.S. PALs). Any reluctance on the part of the launch crew or the chain of command to provide the highly secret word would simply mean that if the missile were launched without authorization or by accident, it would be visibly destroyed immediately upon launch. There has been a general reluctance to consider active meas- ures for destruct after launch which would involve radio transmission to the missiles, but this passive destruct af- ter launch system is much less controversial, evoking the response, "How does that differ from a PAL?" Nevertheless, it can be added in parallel to existing control systems and be made both user-friendly and secure. I emphasize in brief paper(4) which I ask be included in the record as part of my testimony, that if a small ICBM threat should develop, an important element of the defensive system is the same infrared warning satellites ("DSP") that were revealed by the Administration during Desert Storm to have detected every single SCUD launched from Iraq at Israel or at Saudi Arabia. An isolated ICBM reentry vehicle could then be intercepted in mid-course on its way to the conti- nental United States by one or two interceptors launched from the single site at Grand Forks, ND, permitted by the 1972 ABM Treaty. This would be simpler and more effective than to use space-based interceptors, "Brilliant Pebbles", as prescribed by SDIO in their GPALS proposal. Of course, if the problem of mid-course discrimination can't be solved, neither the Brilliant Pebbles nor the ground-based interceptors can do the job of mid-course intercept. More effective, in any case, is pre-boost phase intercept, which is easier against an ICBM than against a SCUD, because the ICBM is much larger and less readily concealed. As was emphasized also in previous testimony, even an effec- tive missile defense of the United States would not signif- icantly reduce the overall nuclear threat, which is far more likely to arrive in the form of smuggled or air-delivered nuclear weapons than via ICBM. In a written response for the record, Harold Brown explained that even a ground-based U.S. missile-defense system could prove destabilizing if the other side ("however misplaced the expectation") came to believe that with the reduced strategic offensive forces a potential for expansion ("breakout") of the ABM system "might be able to intercept all or almost all of a force de- graded by a preemptive strike. That is a recipe for insta- bility." The defense of deployed troops or friendly capitals against non-nuclear missiles of theater range, as analyzed in my pa- per, can better be done by ground-based interceptors, again, than by space-based. This is even clearer for the theater ballistic missiles (TBM) because the boost phase is so brief that space-based weapons have no time to reach the target, which in any case remains in the dense atmosphere during rocket motor firing, and the maximum height of the arc (apogee) can be held below 100 km so that a Brilliant Pebble would have little capability to intercept. A ground-based missile system based on Patriot, but modified to capitalize on the important cueing available from DSP could do a pretty good job against an attack that does not involve decoys or countermeasures, even as simple as those accidentally pro- vided by the breakup of the stretched SCUDs launched by Iraq. Other interceptors under development by the U.S. Army, e.g., THAAD (Theater High Altitude Air Defense), could cover a larger area. However, we should be realistic in our expectations, since it is easy enough to improve the effec- tiveness of the TBM as a terror weapon by dividing its ex- plosive, chemical, or even biological payload into 10-40 bomblets, which would be separated at the end of powered flight. No active defense has been proposed that would in any way effectively engage the bomblets; and bomblets would increase the effectiveness of CW and BW weapons and of antipersonnel explosives. A BAN ON ANTI-SATELLITE TESTS AND WEAPONS IN SPACE. U.S. space satellites conduct vital navigation, observation, and communication tasks, the value of which was strikingly dem- onstrated in Desert Storm. Some of this information can be made available to the United Nations and to the Commonwealth of Independent States to build confidence and security; in- deed, on February 18 Secretary of State James R. Baker and Russian Foreign Minister Andrei Kozyrev announced talks aimed at establishing a joint Early-Warning Center. The emerging Open Skies regime will allow nations or groups of nations without such satellite capabilities to obtain infor- mation by aircraft overflight, in many cases, which, espe- cially for limited areas, can provide better information than that available from satellites. Space-based sensors such as DSP also have an important role in defense, but the one capability that space-based weaponry would clearly have is to destroy satellites. Our valuable satellites are few and vulnerable. They would benefit from legal protection, which would then legitimize enforcement action. I believe that U.S. security would be enhanced by a ban on all space weapons (not sensors), and by a ban on test and use of antisatellite weapons (ASAT), whether launched from land, sea, or air. Nations that are denied the pros- pect of developing nuclear weapons could perfectly well rise to the technological challenge of perfecting ASAT, partic- ularly against satellites in low earth orbit. The United States does not need the spread of such capabilities. COOPERATIVE SECURITY MEASURES. The changes in political structure and military capability have been so abrupt and so severe that we have been caught unprepared to exploit them for our mutual benefit. Of course, START itself and the reciprocated Presidential initiatives are examples of coop- erative security measures, but so is targeted aid to the Commonwealth for the transportation and storage of warheads slated for destruction. Installation of passive destruct- after-launch systems in Commonwealth strategic weapons would respond to U.S. security concerns, and, even after some sharing of the cost, would save the U.S. time and much money in resolving our concern about accidental or unauthorized launch. A new opportunity for cooperative security would enhance the confidence of each side that the ICBMs of the other side re- main in their silos. This could be achieved by mounting small, specialized radio transmitters on each silo cover, generating an unpredictable code that would be relayed by a normal communications satellite to a command center of the other side or even to the United Nations. Arranged so that the communication would cease if the transmitter (or silo cover) were moved, the system would give continuing assur- ance that ICBMs were not on the way. This would not be, in truth, a warning system but a confidence-building measure, since the transmitter could readily be destroyed or disa- bled, thus denying any possibility of report of actual launch. But if a signal stopped, and especially if many stopped at once, the ICBM owner would be highly motivated to explain how this had happened, in order to avoid the presumption of launch. Another cooperative measure would have the greatest impor- tance in achieving the goals of the Missile Defense Act of 1991,namely extending to all nations the U.S.-Soviet INF Treaty. This bans the possession, worldwide, by the U.S. or the Soviet Union (and now by any of the successor Republics) of ground-launched missiles of range 500-5500 km, whether cruise or ballistic, and whether nuclear or conventionally armed. Obviously, U.S. and Commonwealth security would be enhanced if this treaty were extended universally, and, un- like active defense, a treaty could eliminate the missile- delivered bomblet threat as readily as it could handle simple missiles. It is truly a wonder that the U.S., the Soviet Union, and Iraq are the only nations in the world le- gally barred from possessing land-based missiles of short and intermediate range (INF); if the ban could be extended to almost all nations, there would again be a basis for sanctions against nations that refused to join or to comply. STILL ROOM FOR IMAGINATION AND GUTS. President Bush's Sep- tember 27 initiative seized the moment and set into motion an extremely significant process. So in this "revolution- ary" new world of arms control, as ACDA Director Ronald F. Lehman, II, called it in his testimony, new opportunities arise. Some are old ideas whose time has come. According to Mr. Lehman: "Just as the political changes that you mentioned have opened up what some of us have jokingly called the whistle blower approach to verifica- tion, that is to say, the access to people who may want to reveal when their countries are not com- plying with international obligations, so openness measures such as open skies have given us new op- portunities." Benefitting from such information in Iraq (and in the enforcement of IRS regulations and various domestic laws), we have yet to maximize our benefit from verification whis- tle blowers. Future treaties and agreements should have a clause requiring their wide publication in summary form, with a clear statement by the host government that it is the duty of each person having knowledge of any violation of the agreement by his or her government to communicate that know- ledge to the relevant international organ. Enabling legis- lation might be required, and thought given to mechanisms to minimize retaliation; but that is why we have an ACDA and others inside and outside government thinking of new or newly feasible approaches. The United Nations mechanism was very useful to the nations of the world in dealing with Iraq in 1991. We will place more dependence on the UN and the IAEA in countering nuclear proliferation. Serious thought should be given to a UN role in providing positive security guarantees to nations eschewing nuclear weaponry; perhaps with forces committed to UN use by the nuclear powers. SUMMARY START should be ratified and implemented as soon as possi- ble. The post-START era is well under way with the initi- atives set in motion by President Bush September 27, 1991, and disablement and guarded storage of many thousands of nu- clear warheads must be our urgent near-term objective. Post-START forces should have an early goal of 3000 actual strategic warheads on each side, plus perhaps 400 air- delivered tactical weapons. As soon as other nuclear powers see merit in reductions, the U.S. and Commonwealth should move each to 1000 single-warhead weapons. The threat of ac- cidental or unauthorized launch should be handled by instal- lation of a passive destruct-after-launch system rather than by active defenses, and theater-range missiles should be eliminated by extending to all nations the INF Treaty which binds the U.S. and the successor states of the Soviet Union. ---------------- 1 Testimony to the Senate Armed Services Committee, Febru- ary 5, 1992, by Reginald Bartholomew and by Stephen J. Hadley. 2 Testimony of Robert M. Gates to the Senate Government Affairs Committee, January 15, 1992. 3 S. Frankel, Testimony to the Failsafe and Risk Reduction Committee of the Department of Defense, November 5, 1991. 4 "Space-Based Defense Against Ballistic Missiles," by R.L. Garwin, presented to the American Association for the Advancement of Science in Chicago February 8, 1992.