206. Memorandum of Conversation1


  • Summary of Meeting with the President on CTB Issues


  • President Jimmy Carter
  • Harold Brown, Secretary of Defense
  • James Schlesinger, Secretary of Energy
  • David Aaron, Deputy Assistant to the President for National Security Affairs
  • Stuart Eisenstat, Assistant to the President for Domestic Affairs (first 10 minutes)
  • Dr. Frank Press, Director of the Office of Science and Technology Policy
  • John Marcum, Senior Adviser, Office of Science and Technology Policy
  • Harold Agnew, Director, Los Alamos Nuclear Laboratory
  • Roger Batzel, Director, Livermore Nuclear Laboratory

Harold Agnew began by saying that he understood that it might be good if he or Roger Batzel ran through a brief explanation of how nu[Page 498]clear weapons actually worked. The President agreed saying he had read some simple text books but wanted a more complete understanding.

Harold Agnew presented a chart2 of a simple two-stage device to the President and explained its design features. He noted that we had made tremendous progress in nuclear weapons technology and illustrated this by pointing out that the current Trident warhead had a yield [less than 1 line not declassified] with a weight [less than 1 line not declassified] compared to a yield [less than 1 line not declassified] of the Hiroshima device with a weight of [less than 1 line not declassified] He said this represented a factor of more [less than 1 line not declassified] in yield to weight ratio. He observed that this tremendous gain was made through very sophisticated technology. One major step was that instead of using a solid glob of plutonium, we now use a [less than 1 line not declassified] With this design we can get [less than 1 line not declassified] in the primary or first stage, but an even more significant step was in using tritium gas for “boosting” the primary. This boosting process produces [less than 1 line not declassified] than the [less than 1 line not declassified] and increases the primary yield [less than 1 line not declassified] causes the secondary stage of the weapon to ignite.

Harold Agnew noted that over the years nuclear weapon designers had been under considerable pressure to develop designs that would use less fissionable material, which could be in short supply for civil purposes, and that would be smaller, lighter and safer. In explaining what the two-stage primary act really does, he said that the Trident primary has a yield [less than 1 line not declassified] the secondary [less than 1 line not declassified]. The important consideration is that the primary must produce [less than 1 line not declassified] there would be no ignition from the secondary.

He then showed a picture of the Sprint ABM warhead3 noting that this was the first neutron bomb that had been developed. He described its features as [2½ lines not declassified] This device was tested and achieved [less than 1 line not declassified] He then explained that at a later point because of fratricide concerns—e.g., concern that the warhead might be disabled [less than 1 line not declassified] another Sprint warhead that had been detonated earlier, it had been decided to add a hardened layer [3 lines not declassified] He noted that this wasn’t a complete fizzle but was less [less than 1 line not declassified] necessary to ignite the secondary stage. He asserted that if we had stockpiled the modified device without testing we wouldn’t have known about the problem; stockpile surveillance was essential but could not determine the yield of the device.

[Page 499]

The President asked what kind of stockpile surveillance method was used. Agnew responded that all kinds of statistical sampling techniques were employed, weapons were disassembled occasionally, and the components were examined. He noted that the weapons were designed to last 20 years, but that any time the materials used in manufacturing components are changed one can get problems. He pointed out that the devices we had developed were very delicate and that in his view it would not be prudent to consider maintaining the stockpile indefinitely without testing.

The President asked what Agnew meant by indefinitely. Agnew responded that in an example which Roger Batzel would describe, a problem had been experienced within four years. Roger Batzel explained that [3½ lines not declassified]

The President asked whether the problem had occurred after development of the warhead. Batzel replied that it had actually been after deployment of the warhead and had been discovered through surveillance of warheads deployed in the fleet. The problem had been a mechanical safety device which was intended to insure that the weapon would not produce a significant nuclear yield if the high explosive (HE) was detonated accidentally.

The President said that he understood that the problem was not identified in an explosion but in an inspection of the safety device. Batzel agreed and stated that we had subsequently replaced the primary on that device. He pointed out that we have a similar problem now with the primary of the Poseidon warhead. He pointed out that the Poseidon produced a nuclear yield [less than 1 line not declassified] packaged in a re-entry vehicle weighing [less than 1 line not declassified] He explained that although primaries used to have hundreds of detonators distributed over the surface of the HE in order to generate a spherical implosion wave, [1 line not declassified] In the Poseidon primary a new HE material had been used which exuded a liquid which corroded [less than 1 line not declassified] that with time, no yield would have resulted. Fortunately, during the process of development, an alternative design using different HE material had been fully tested and we were able to simply substitute this primary for the defective one. Otherwise further nuclear testing would have been required.

Roger Batzel also argued that US designers in responding to military requirements had made remarkable achievements in minimizing weight and maximizing yield of warheads. [3 lines not declassified] Agnew interjected that without the one-point safety requirement we would not be so close to the edge. Batzel agreed, adding that a factor of two increase in weight would also avoid some problems. He also remarked that while these were remarkable accomplishments we might [Page 500] have pushed the technology too hard and created problems for ourselves in the current context.

Harold Agnew said that people could say that we haven’t been testing stockpiled weapons for reliability for years, why now? He said that they didn’t realize the benefits to stockpile maintenance that are achieved from testing new weapons and continuing to produce nuclear weapon materials and components. He observed that many devices which are being tested in new warhead development programs use the same primaries which are in the old stockpiled weapons.

The President asked whether the Soviets had more reliable warheads than the US. Harold Brown answered that in general their warheads were believed to be heavier, somewhat more roughly constructed and less sensitive to deterioration. Roger Batzel agreed and pointed out that they had not had the same constraints on weight and size, and appeared to have developed less sensitive warheads.

The President asked how long we have had small primaries. Roger Batzel responded nearly 20 years and after thinking about Polaris and Poseidon, said it was really about 10 years. Harold Agnew pointed out that development of smaller primaries had really been a result of the pressure of MIRVing of ICBMs and SLBMs.

The President asked what had been our experience with regard to correcting problems in inventory. Roger Batzel replied that in the early 1960s a vulnerability test of one of our ICBM RVs had been carried out. The President asked if this was for fratricide purposes. Batzel said yes but added that it was primarily to check for vulnerability to Soviet ABMs. [3 lines not declassified] Harold Brown pointed out that [2 lines not declassified]

Harold Agnew stated that another example was with our B–25 air-to-air warhead which consisted of [less than 1 line not declassified] The plutonium at that time was manufactured in Hanford and Rocky Flats—one making [less than 1 line not declassified]. After the weapons were stockpiled, people began to notice [less than 1 line not declassified] After disassembling them it was discovered [less than 1 line not declassified] This was due to very slight differences in the manufacturing processes at Hanford and Rocky Flats so that the plutonium [less than 1 line not declassified] might shrink slightly while that [less than 1 line not declassified] expanded.

Agnew said that another example where the warhead did not perform as expected was the Talos/Terrier as mentioned earlier, [less than 1 line not declassified] As a result of these problems he felt that we wouldn’t be able to certify warheads if the materials were changed in any way without testing.

Frank Press commented out that he had brought a panel of experts together to look into this issue. His panel had included the current labo[Page 501]ratory directors, former laboratory directors, such as Herbert York, and other nuclear weapons experts. He pointed out that the laboratory directors concerns were as they had been stated today but the other members had a different view and felt that reliability could be maintained for at least five years without testing, and that this had been our basis in recommending the five year duration approach to the President.

The President commented to the laboratory directors that one concern he had is whether they were able to maintain their objectivity on this issue in view of their desire for preservation of the laboratories. He said he was trying to put a lid on production and development of all new nuclear weapons and that an important element of this is to put a cap on testing. He stated that we must maintain reliability of our nuclear weapons but that he would need as much flexibility from the laboratory directors as possible to accomplish his overall objectives. He pointed out that without their support a test ban would experience serious difficulties with Congress. He noted that his own advice on the issue was conflicting in that he had decided to ban testing for five years while preserving the right to resume testing. He continued that in his view a threshold test ban would circumvent the basic purpose of the treaty. He wanted to make clear he was not referring to a few pounds but to hundreds of pounds or a kiloton. He pointed out that had discussed this issue earlier in the day with Prime Minister Desai of India,4 and Desai had said that a 5KT level would open the flood gates to proliferation. He said he was eager to learn from the laboratory directors but expected them to support his objectives as much as possible.

Roger Batzel said that they were trying to be responsive but the problem was that they didn’t know what problems would arise in 2, 3, 4, or 5 years.

The President then asked what the yield range was for our present primaries. Harold Agnew replied that they ranged [less than 1 line not declassified] the Titan down to [less than 1 line not declassified] some of our other nuclear systems. Harold Brown pointed out that the Soviets are believed to have larger primaries involving heavier, more rugged designs which are probably less sensitive to deterioration. Roger Batzel said that although some of our primaries had higher yields, testing [less than 1 line not declassified] enable us to recertify these primaries when problems arose. He said the laboratories have tried to be responsive to the President’s objectives by holding the yield as low as possible, [less than 1 line not declassified] reproducibility is a real problem that there was se[Page 502]rious doubt that testing [less than 1 line not declassified] be useful. The difficulty was that it is necessary to get significant boosting in order to insure the primary worked properly. He said that to date we haven’t resolved that testing [less than 1 line not declassified] was useful.

Harold Agnew elaborated on this point, explaining that we have tested [less than 1 line not declassified] adding small amounts of boosting gas but the yield varied widely and was unpredictable as the boosting gas was added. There is no consistency in the results until the yield gets [less than 1 line not declassified]

The President asked whether there was any statistical evidence on the difference in warheads between 3, 4, and 5 years, noting that he gathered there was a difference in view here. Harold Agnew responded that we expected all our weapons to last 20 years but as noted earlier in some cases, such as Polaris, problems had been discovered in just four years. The President pointed out they had not discovered it through testing. Harold Agnew agreed and said they had discovered it by surveillance.

The President asked whether they had ever discovered in a stockpile warhead any physical deterioration by nuclear testing. Agnew replied that the [less than 1 line not declassified] when taken from the stockpile and tested had produced an [less than 1 line not declassified] However, it had never been tested before since it had been developed and stockpiled while the moratorium was in effect. Harold Brown pointed out that was a result of not testing the warhead rather than of deterioration after the warhead had been properly tested and stockpiled.

Frank Press , asking that the laboratory directors correct him if he were wrong, pointed out that it was his understanding that most of our warheads had worked properly the first time they were tested. Harold Agnew responded [3 lines not declassified] Frank Press agreed that such problems had occurred but stressed that most of the time the devices had performed well the first time they were tested. Harold Agnew acknowledged that this was true and said that was an argument frequently used by some of Frank Press’ colleagues. In fact, 80% or so did work the first time but the problem was that this might not include Minuteman or Trident or some other important warheads.

David Aaron asked whether there hadn’t been enough testing of enough designs over the years to get a good basis to predict how the weapons would perform. The President asked whether he was referring to new designs. Harold Brown said that that was not the issue. The question was whether there would be deterioration in the stockpiled weapons within five years. If so, then in his view, Frank Press’ point was right—we could redesign and rebuild the weapon and have confidence that it would probably work.

[Page 503]

David Aaron asked Frank Press whether his panel had looked into the previous record on the problems that had arisen in the stockpile. Frank Press responded that his group had reviewed the record and it was essentially just as the laboratory directors had said today. In his view, the important point was that if the weapons were rebuilt the same way as originally, they would have high confidence in their performance.

The President asked whether stockpiled warheads were routinely tested. Roger Batzel replied they were not and Harold Agnew noted that we do have proof tests of each type of warhead before it is stockpiled which certifies performance of all warheads of a type before they are put into a stockpile. The President said this was good, otherwise we would have to test every weapon. Harold Agnew returning to the question, said that the primaries of old warheads were sometimes tested in development of new weapons and sometimes the entire stockpiled warhead would be tested for other reasons, such as vulnerability, but not routinely.

Harold Brown offered to more clearly explain the issue, he said that once the final proof test is completed, the weapons would be stockpiled and subjected to surveillance procedures, but would probably not be tested again unless for some other purpose such as Harold Agnew had mentioned. Roger Batzel pointed out that some very peculiar problems had arisen over the years even in the nuclear components themselves. [2½ lines not declassified]

The President asked Jim Schlesinger whether he had any thing he wanted to mention in this discussion. Schlesinger responded that there were two points he wanted to raise. The first was that this was not a typical statistical problem and that we are trying to determine the probability of a unique unpredictable event and we don’t have any idea when or whether it will occur. In discussing his second point he presented a chart showing the utility of testing at various yield levels—from a few pounds to ten tons would be useful for one point safety and hydronuclear testing, from ten tons to 10 KT for reliability testing, and from 10 KT to 150 KT for development of new weapons.

Schlesinger explained that the main point was that [less than 1 line not declassified] we could meet our needs but that lower yield levels did not give us the confidence that the primary will drive the secondary. He said that left to their own devices the laboratories would prefer to go on designing new nuclear weapons. In accomodating the President’s desires they had given up on that but they still had the responsibility of certifying stockpiled warheads. As he pointed out this would require testing [less than 1 line not declassified]

The President noted they didn’t routinely test stockpile weapons. He then asked for the typical reliability of the launchers for these war[Page 504]heads. John Marcum replied they were about 80%, Harold Brown said yes, 75–80% but pointed out the real difference was that for the launchers they didn’t have to worry, there was a 25% chance that none of them would work.

The President said it was his understanding that there would be a provision that if he or Brezhnev had a pressing national need they would be able to withdraw from the treaty. Harold Brown agreed, saying there would be a “supreme national interest” withdrawal clause, but it would be very difficult politically to exercise that clause since it would abrogate the treaty. Further, there would be undoubtedly fierce interagency fights over whether the particular problem was serious enough to justify withdrawal.

David Aaron pointed out that it might be possible to define a flexible “supreme interest” clause that would let either side resume testing for national security reasons without collapse of the treaty. The President agreed that this might be possible.

Jim Schlesinger interjected that an alternative would be to have a small quota of tests below 5 KT and that this would minimize the difficulty associated with the threshold. The President noted that from what Harold Brown said we would be able to test if really necessary and asked how long it would take us to do so if a problem arose.

Roger Batzel replied that this would take from 6 to 9 months depending on the level of readiness provided for in the Safeguards Plan. The President asked how long the moratorium had lasted. Harold Brown said it lasted about 2½ years from the summer of 1958 to early 1961. Harold Agnew said that a lot of people thought he was paranoid but that he had never gotten over this. He presented a chart5 showing that the Soviets have resumed testing immediately with a massive and well prepared program, whereas it had taken us an extended period to carry out the first test. Frank Press told Harold Agnew that we really needed his and Roger Batzel’s help in designing a Safeguard Plan which would provide the right level of readiness and help keep our weapons group together.

Harold Agnew said that he understood there could be concern as the President mentioned with regard to whether the laboratories have a vested interest in testing. He assured the President that their only concern was to maintain our nuclear deterrent. He said that the laboratories constitute a valuable resource and they happen to think they are smarter than anyone else and could do a better job in solving most new problems. He commented that their staff spent 40% of their time on [Page 505] weapons and 60% on energy activities and they would supply the technical effort wherever the President wanted it.

The President asked what else they would do in maintaining the stockpile. Harold Agnew said they had been examining this question and, in his view, their present surveillance procedures were adequate, although they could, of course, increase this program if it made anyone feel better. He pointed out they used to test one out of fifty weapons each year, but now did much less of this sort of testing.

Frank Press asked the lab directors whether they saw any real hope for reliability testing at the 100 KT level. Both replied they didn’t think this was likely. Harold Agnew pointed out that [2½ lines not declassified] Harold Brown asked [2½ lines not declassified] Harold Agnew concluded that [1 line not declassified]

The President, noting the particular names of some of the tests, asked Agnew for the origin of those names. Harold Agnew responded that many of them were place names from New Mexico. James Schlesinger pointed out that there was a great deal of empiricism in nuclear weapons, that in theory, it always appeared they would work well.

The President said his only remaining concern after this discussion was the need to consider the relative effect of no testing on the Soviets. As he understood there was no real effect, except for the massive weight of their warheads. Harold Agnew said there was another point in that Soviet missiles have much greater throw weight, which meant that if a problem arose they could simply put another warhead on, which might be less optimal in terms of weight or size, more easily than we could. Harold Brown pointed out that since the size of their warheads were larger, they might have to test at a higher yield than 3–5 KT. On the other hand, since their warheads were less sensitive, they might not have to test at all.

James Schlesinger [3 lines not declassified] Harold Brown [2 lines not declassified] John Marcum [1 line not declassified]

Roger Batzel [5 lines not declassified]

The President asked Batzel [1½ lines not declassified] Batzel [1½ lines not declassified] Harold Brown [1½ lines not declassified]

The President said he had a question for Frank Press, that he had just heard we didn’t have sensing devices monitoring the treaty. Frank Press assured him that we did have these devices for use in internal seismic stations and they had been installed in bore holes in New Mexico for testing. Although we had not completed packaging of these sensors, we should be ready to begin installation in the Soviet Union as soon as the treaty entered into force. He estimated it might take two years to install a complete network of single stations and perhaps three years for arrays.

[Page 506]

The President asked how many seismic stations we were talking about proposing in the negotiations. Frank Press replied that our analysis indicated that a network of about 17 single stations would be roughly comparable to about five arrays. James Schlesinger responded we were thinking of proposing 12–15 single stations and this would get us down to a threshold of about [less than 1 line not declassified]

Harold Brown and Frank Press objected noting that the real threshold would be lower when other national means were taken into account. Frank Press said we should be precise on this issue, that with that kind of network we would detect down to 2/10 of a KT in many cases. He noted that in seismic regions detection would be equivalent to identification of the event and that in seismic regions, we were looking at very remote locations so that any suspicious activity from satellite photos would help identify the event. Other national means could also help identify problems, and in general the identification threshold should not be more than two times the detection threshold, or about 4/10 of a KT in rock. He said this should be increased by a factor of five in looser material and conceivably by a factor of 10, which would get to the level James Schlesinger had mentioned, with dry alluvium. [7 lines not declassified] Harold Agnew inserted that the Soviets were doing this right now. Frank Press noted that we had also conducted tests in cavities.

The President [1 line not declassified] Harold Agnew [1 line not declassified] John Marcum [4½ lines not declassified]

James Schlesinger showed the President another chart6 illustrating the problem of unidentified events and said that even with the seismic network we had in mind there could be 38 or so unidentified events annually. He said we would be pretty comfortable in verifying at the 5 KT level. He noted that seismic arrays were very expensive and that it was probably not worth driving this down to 4 KT, although it might give us more support on the Hill. He asserted that it was the combination of our inability to verify or certify that would give rise to serious domestic political problems.

Harold Brown said these were unrelated problems and that as noted earlier the Soviets might not need to test at all. He felt the verification problem was a consequence of our previous statements that any agreement we negotiated must be highly verifiable. He felt the Soviets were unlikely to cheat under the five year approach, but said that James Schlesinger was right in that verification would be a political problem. There was no doubt that a 5 KT threshold or a small quota would be better in terms of a Senate ratification effort.

[Page 507]

The President asked if we had a real low threshold for a short period, he emphasized that he meant very short, of perhaps six months, what would the lab directors want to test? Harold Agnew responded that if other needs had been met they would want to use their quota to test stockpiled warheads for reliability purposes. The President asked how they would view three years, maybe with a small quota. Roger Batzel replied that a duration of three years would be much easier than five years in maintaining laboratory capabilities. Harold Brown noted we had gotten through 2½ years during the moratorium. Roger Batzel agreed but said it was starting to hurt and that more than three years would be particularly tough.

Frank Press commented that offering opportunities to work in laser fusion and other related areas could be helpful in retaining scientists. Harold Agnew agreed but argued that the best people would switch into other areas on a permanent basis. In time we would lose our good people and have little confidence in our stockpile, but he guessed that presumably this was consistent with our long range objective.

David Aaron said in his view it wouldn’t help to have a short period of testing unless a problem had been identified which needed to be corrected. Harold Agnew agreed and pointed out that what was really needed was a 3–5 KT threshold to cover any problems that would arise and that this would be a major step forward in restraining testing.

The President asked what the effect of a CTB would be on SALT. Harold Brown responded that SALT constraints were not applicable to warheads. Roger Batzel commented there was a relationship in that the CTB would constrain our ability to provide warheads for new missile systems.

Harold Brown said that if we don’t test for three years and that, in that time, deterioration had not occurred, then the same problems would be experienced after three years as now. The President said we could build all new warheads at that time for critical systems. He asked if SALT II permitted new missiles what would be the effect of a CTB? Harold Brown said there would be little effect, since if we were going to resume testing after 3–5 years we would just design new warheads and test them at that time.

John Marcum pointed out that it was his understanding that prior to the threshold test ban we had fully tested new warheads which were intended to meet the future needs of the M–X, TRIDENT II and cruise missile systems. Harold Agnew agreed that this was the case and said we had a family of tested warheads for use and planned to use these in developing new systems, since we would be unable to develop new strategic warheads even at 3–5 KT. The President commented then the 5 KT threshold would not really help in this regard. Harold Brown said probably not although new tactical warheads might be developed. Roger [Page 508] Batzel commented that in testing these new warhead designs the laboratories had tried to anticipate future strategic needs.

The President asked what the yield was on the largest device we had tested. Harold Brown replied [less than 1 line not declassified] Harold Agnew said it was more [less than 1 line not declassified] and for a very dirty device (lots of fissionable material) the yield could have been as high [less than 1 line not declassified]

David Aaron asked whether from the laboratory perspective it would be better to have a small quota for reliability testing each year or to have unlimited testing after five years. After having the question repeated, Roger Batzel responded that they would probably prefer a small number of tests each year.

The President told David Aaron that with either a quota or threshold he thought we would lose our non-proliferation impact and the other political benefits we were seeking. In his view, a 5 KT level would be high enough to permit further proliferation in other coun-tries and might be just fine for India. Harold Brown said he agreed completely.

Harold Agnew asked whether 2 KT would matter. The President said he thought so, that the essence of our position is that for a period of time we would not test at all and then could resume testing.

The President said he wanted to make clear that he did not share all the laboratory director’s concerns. He said he wasn’t sure at all that he shared their concerns about problems arising within five years, but said he might not be qualified to judge. Harold Brown said this was a judgmental issue and the President’s opinion was as good as anyone’s.

The President asked whether there was any way to make the warheads less sensitive, possibly by adding more tritium. Harold Brown responded that we might make them less sensitive by relaxing the one-point safety criteria, explaining that this requirement meant that an accidental detonation of the HE at one point should not result in a significant nuclear yield and that this requirement had required us to minimize the plutonium in the warhead.

Roger Batzel said they had examined all these ideas, that to redesign the stockpile would take a lot of time and could have some disadvantages, and wouldn’t really help much. He said that at the expense of more weight we could double the HE. Harold Agnew noted we could also add plutonium to make the primary hotter so it would be more likely to achieve an effective tritium burn.

The President said he had to leave for another appointment and that the meeting was very useful. The meeting adjourned at 3:30 p.m.

  1. Source: Carter Library, National Security Council, Institutional Files, Box 95, SCC 084, CTB Verification, Seismic Station Network: 6/12/78. Secret. The meeting took place in the White House Cabinet Room.
  2. Attached but not printed.
  3. Not attached.
  4. According to the President’s Daily Diary, Carter met with Desai in the Oval Office from 11:01 to 11:26 a.m. (Carter Library, Presidential Materials) A record of their meeting is scheduled for publication in Foreign Relations, 1977–1980, vol. XIX, South Asia.
  5. Attached but not printed.
  6. Attached but not printed.