53. Memorandum From Jan Lodal of the National Security Council Staff and the Counselor of the Department of State (Sonnenfeldt) to Secretary of State Kissinger1


  • SALT—Flexibility in the MIRV Throw Weight Approach

In light of Soviet reactions thus far to MIRV throw weight concept, this memo explains how this approach can meet possible Soviet concerns while advancing our security interests.

Our Present ICBM MIRV Throw Weight Proposal

The proposal you suggested to Gromyko would allow both sides 1.3 million pounds ICBM MIRV throw weight.2 The Soviets would get [Page 192] about 240 of their SS–19 medium ICBMs. For strategic and verification reasons, the other Soviet MIRV and ICBMs (SS–17 and 18) would be banned. The U.S. could have about 550 MM III. There could be no changes to silos not containing MIRV ICBMs.

Under our proposal, we would have the advantage in MIRV launchers and RVs while the Soviets would have the advantage in non-MIRV launchers and total throw weight.

The Soviets may have several concerns with this approach.

—They clearly want to limit our SLBM MIRV program.

—They probably want to deploy more than 1.3 million pounds of MIRVed ICBM throw weight.

—They almost certainly do not want to ban the SS–17 and 18 (they may want to deploy some of each missile).

This memo deals with these issues in two parts. Section I discusses how our SLBMs might be folded in to provide higher MIRV throw weight levels yet give both sides a more equitable position. Section II addresses the issues involved in the ICBM sublimit—in particular banning the SS–17 and 18 and the collateral control needed if they are to be permitted.

I. SLBMs and the Level of MIRV Throw Weight

In SALT I, our negotiating strategy focused on the linkage between ABMs and offensive systems. We exploited the leverage of our ongoing ABM program to set a ceiling on the Soviet offensive buildup in addition to halting ABMs.

Our basic objective in SALT II is to achieve significant limits on the Soviets’ new ICBM programs. These new programs, if unconstrained by SALT, will translate a strategically meaningless Soviet ICBM number and throw weight advantage into a potent strategic advantage which could open up counterforce options for the Soviets that would be unavailable to the U.S. and destabilizing.

While Schlesinger talks bravely of starting up new U.S. ICBM programs and new counterforce RVs to match the Soviets, in fact, the real source of our bargaining leverage in SALT II is our dynamic SLBM program (e.g., the Trident missiles and boat) and probably the B–1 and ASMs.

Moreover, there is strategic connection in this linkage. If we cannot control Soviet counterforce development, we will want the most modern and advanced SLBM and bomber forces. But if worthwhile controls are possible, we can afford to give up at least some of the most advanced SLBM options (e.g., Trident II missile).

In sum, how SLBMs are folded into an agreement is not a tactical question but is a focal point of our SALT strategy, and probably your strongest card.

[Page 193]

The U.S. Program—Trident

As you are aware, the Trident program has three basic elements:

—The Trident I missile which will carry a Poseidon-size payload to a range of about 4,000–4,500 n.m. It can be deployed in either Poseidon or Trident boats and thus greatly improve their survivability.

—The Trident submarine itself is about twice the size of a Poseidon submarine and will initially carry the Trident I missile when it is deployed in the late 1970s.

—The Trident II missile is much larger than Trident I and will fully utilize the large tube volume available in the big Trident boat. It can only fit in the Trident boat. It can come in two versions: a 6,500 mile Poseidon/Trident I size payload of 4,000 pounds, or a high throw weight 6,000 pound payload with only 3,000 mile range.

—An alternative to Trident submarine is in the new budget; a smaller, cheaper, but equally effective boat based on the so-called Narwhal natural circulation reactor. The Narwhal alternative is a Poseidon size submarine which would carry Poseidon size tubes (i.e., it could carry Trident I but not Trident II).

The main point is that despite the large political investment and risk, the Trident boat and Trident II missiles are expendable at SALT. But, we must play the Trident card reasonably soon. Now that Congress has got wind of a serious competitor, the Trident boat may well be killed.

The Soviet SLBM Program

While we have very little hard intelligence, CIA’s best estimate is that they have MIRV missiles under development for both the Y and D Class. This information is based on [less than 1 line not declassified] modifications underway at various SLBM facilities. We have no idea what the throw weight of Soviet SLBMs might be and it is difficult to assess how a MIRV throw weight limit would impact the Soviet SLBM program.

Flexibility Under the MIRV Throw Weight Approach

The Soviets are likely to push our basic ICBM MIRV throw weight proposal in two directions:

—First, the Soviets probably want to deploy more than 1.3 million pounds of ICBM MIRV throw weight and will try to drive up the ICBM level.

—Second, the Soviets undoubtedly want to include SLBMs in such an agreement and will try and stop some aspect of Trident as well.

The key is the linkage between the two: how low we limit ICBM MIRV throw weight will affect the stringency of limits on our SLBMs.

[Page 194]

Our options are basically as follows:

OPTION 1. If the Soviets accept the 1.3 million pound ICBM MIRV throw weight limit, we should be willing to hold our SLBM MIRVs to our current level of 500 Poseidons, i.e., 2.1 million pounds. We could deploy Trident I missile since it has the same throw weight as Poseidon. Deployment of the Trident or Narwhal boat, if permitted at all, would have to be as a replacement for Poseidon. The combined ICBM/SLBM total would be 3.4 million pounds. The Soviets would get 240–300 SS–19/17s and could MIRV both the Y and D class submarines.

The ICBM/SLBM forces of the two sides would be:

United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
550 MM III 1.3 240–300 SS–17/19 1.2–1.6
496 Poseidon/Trident I 2.1 950 SLBM MIRV 1.8–2.2
MIRV Subtotal 1046 3.4 1190–1250 3.4
United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
450 MM II 1.0 790–730 SS–11 1.8
54 Titans .4 60 SS–13/16 .1
160 Polaris .2 300 SS–9 3.9
UnMIRVed Subtotal 664 1.6 1150–1090 5.8
Total Launchers 1710 2340

OPTION 2. If the Soviets want higher ICBM MIRV throw weight limit, at or below 2.4 million pounds (i.e., equivalent to 1000 MM III); we would still limit SLBMs to 2.1 million pounds. However, we would insist on the right to 1000 MM IIIs and on freedom to mix to sea to deploy additional SLBM MIRVs as replacement for Polaris (either Trident or Narwhal) instead of MIRVing additional ICBMs. The combined ICBM/SLBM total would be about 4.5 million pounds.

This would give us theoretical advantage of up to 1000 MM III MIRV launchers and it would allow the Soviets to deploy up to 440/500 SS–17/19s.

[Page 195]

However, if we had an ICBM MIRV throw weight limit at this level, it is more in our interest to deploy the additional MIRV throw weight in our SLBMs instead of MIRVing the rest of our Minuteman force. This would give us an advantage in SLBMs and we could deploy it either in Trident ships, or if we give up the Trident boat, in a Narwhal SLBM carrying Trident I missiles.

The ICBM/SLBM forces of the two sides would be:

United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
550 MM III 1.3 550/440 SS–17/19 2.4
496 Poseidon/Trident I 2.1 950 SLBM MIRV 2.1
240 Narwhal/Trident I 1.0
MIRV Total 1286 4.4 1450–1390 4.5
United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
450 MM II 1.0 530–590 SS–11 1.4
60 SS–13/16 .1
300 SS–9 3.9
UnMirved Total 450 1.0 890–950 5.4
Total Launchers 1736 2340

OPTION 3. If the Soviets want even more ICBM throw weight or no sublimit, we would want to deploy Trident II. We would want to use high throw weight Trident IIs and possibly more MM–IIIs as compensation for Soviet ICBM advantages. We could have up to 3.5 million pounds of SLBM MIRV throw weight and the combined ICBM/SLBM MIRV throw weight level could be as high as 5–6 million pounds. At these levels, the concept of limiting MIRV throw weight becomes somewhat ludicrous and we might be better off to look at some other concept and simply use MIRV throw weight as one measure of the equity of the deal.

[Page 196]

The ICBM/SLBM forces of the two sides would be:

United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
1000 MM–III 2.4 675 SS–17/19 3.7
496 Poseidon/Trident 2.1 950 SLBM MIRV 3.7
240 Trident II 1.4
MIRV Forces 1736 5.9 1625 5.8
United States Soviet
Launchers Throw Weight Launchers Throw Weight
(million lbs) (million lbs)
00 00 355 SS–11 .9
00 00 60 SS–13/16 .1
00 00 300 SS–9 3.9
UnMIRVed Total 00 00 715 4.9
Total Launchers 1736 2340

In conclusion, there are several key points to keep in mind concerning the flexibility of the MIRV throw weight concept and its limitations.

—First, in limiting ICBM MIRVs, we should try for the lowest level. We can accept deployment of the 17 and 19, but it is essential that the Soviets accept the concept of limiting MIRVed ICBMs to specified complexes and to accept a ban on modifying unMIRVed silos. (More on this in Section II below.)

—Second, we should not agree to stop the Trident II missile and Trident ship unless the Soviets stop some of their programs that are of great concern to us; i.e.; the SS–18.

—Third, while we are willing to give up the Trident II missile and ship, we must have the Trident I missiles because of its contribution to SLBM survivability. It is truly a modernization of Poseidon and not a new system. Moreover, the Soviets already have a 4000 n.m. SLBM in their D-class boat.

—Finally, if we give up the Trident boat, we must have the right to replace our Polaris SLBMs and Titans with a new SLBM when they grow [Page 197] old. We would build a Poseidon-size ship carrying Trident I missiles but would not need to do so before the mid-1980s.

II. Limits on ICBMs

Even if the overall outline of an agreement can be worked out, there are some key specifics mainly relating to verification risks that will need to be resolved:

—Which, if any, of the new ICBMs must be banned?

—What other collateral constraints are needed and what are their implications?

To Ban or Not to Ban the SS–17 and 18

The bureaucracy is largely agreed we can monitor deployment of the SS–19 because it won’t fit into existing SS–11 silos. The problem lies with the SS–17 and 18 which many believe should be banned. Our view is that we could tolerate the SS–17 but that we should still try to ban the large SS–18. The argument is as follows:

1. The SS–17. The SS–17 is the new pop-up medium ICBM. Its throw weight is similar to that of the SS–19 (about 5000 pounds), however, it appears to be several feet shorter. We are certain the SS–17 can fit physically in SS–11 silos. And because of its pop-up launch technique, there is a significant risk that SS–17s can fly out of SS–11 silos without deepening or other visible modifications. Thus, from a verification standpoint, the SS–17 is the most troublesome of the new Soviet missiles. Nonetheless, CIA and DIA believe we could have high to moderate confidence of detecting 100 illegal deployments because of the necessary activities associated with this system.

You should bear in mind that the SS–17 also comes in a single RV version with a 5 megaton warhead. If the 17 is permitted, this version must be controlled not only because of cheating possibilities, but because its five megaton single RVs could provide the Soviets with a more potent counterforce weapon than the SS–11.

2. The SS–18. If we do not ban the SS–18 but only control deployment, the intelligence community believes that the Soviets might be able to deploy up to 50–100 SS–18s before we could detect them. Because of the SS–18’s high throw weight, cheating of that magnitude would be very significant from a strategic standpoint. Even if the Soviets are allowed only to test and deploy the SS–18 with single RVs, they will have an enormous MIRV breakout potential, and in any agreement of limited duration we could be under serious pressure as the agreement drew to a close. Banning the SS–18 would lead to gradual obsolescence of the SS–9s and would eliminate their potentially significant role in the strategic equation.

[Page 198]

On balance, we believe that it is more important to try to ban the SS–X–18 and could tolerate deployment of the SS–X–17. The SS–X–18 is the least successful Soviet MIRV ICBM in this test program, but politically and strategically the most important to us. To get such a ban we should be prepared to forego long-range ASMs (which the Air Force doesn’t want anyway) and air-launched ICBMs. Depending on how the overall MIRV throw weight deal works out, we may also stop the Trident submarine which would be additional leverage to stop the SS–X–18.

Collateral Constraints Required

As you well know, verifying such an agreement would not be easy and supplementary provisions are required to give us confidence in verification. The key collateral constraints are:

—MIRVed ICBMs can be deployed only in designated complexes and all ICBMs in those complexes must be counted as MIRVed ICBMs.

—There can not be any changes in the depth, diameter, or configuration of silos at complexes where unMIRVed ICBMs are deployed, nor changes in routine maintenance.

Designated Complexes

The idea of limiting deployment of MIRVed ICBMs to specific complexes is particularly important if we permit the SS–X–17. With such a restriction, CIA has high confidence (DIA moderate) that we could detect 100 illegal SS–X–17s.

A major question is what complexes would be preferred. Sixty silos at two Soviet SS–11 complexes (Derazhnya and Pervomaysk in western USSR) are already compatible with the SS–19 and others there are now being modified. The Soviets have 180 launchers at these two sites. However, the newer modifications do not appear to include deepening so they may be intended for the SS–17 or SS–11 Mod III.

In addition, the Soviets could deploy MIRVed SS–19s–17s at ten other SS–11 sites in the USSR with 50 to 120 SS–11 launchers at each site (any SS–17 deployment would have to be counted as a MIRVed ICBM). A likely candidate as the third site for MIRV deployment is Tatishchevo where a silo modification program is about to get underway and a large III–X silo was recently detected under construction. At Tatishchevo, the Soviets have 120 SS–11s deployed, thus the Soviets could have 300 SS–19s/17s at these three sites. CIA recently examined all of the SS–11 complexes from the standpoint of planned satellite coverage, weather, and other factors, and rated Tatishchevo above the other SS–11 sites as the complex which is best for permitting MIRVed ICBM deployment—(in other words, we can monitor the other locations better for violations).

The U.S. plans to deploy 550 MM–III. Five hundred of these will be deployed at three complexes at Minot, Grand Forks, and Warren Air [Page 199] Force Base. Each of these complexes will have MM–III and no unMIRVed missiles. Thus, if we are dealing with the SS–19/17, we are in a position to suggest that both sides deploy MIRVed ICBMs at only three ICBM sites.

Permitting the SS–X–18 raises special problems if we want to limit areas of deployment. There is at least one new SS–18 (III–F) type silo at every SS–9 complex. Even token deployment at every complex could badly confuse our monitoring capabilities. On the other hand, the silos are not complete and if work were stopped at prohibited sites, the silos need not be destroyed.

Ban on Silo Modification

We are confident that the SS–19 cannot be fired from SS–11 silos without detectable silo modification and there is surmise but not hard proof that some changes would be required for the SS–17 as well—the estimate of its length has increased because we have identified the zero stage and now there is only about five feet of rattle space in the original SS–11 silo. Consequently, the second essential collateral constraint is a ban on changes in the depth, diameter, or configuration and maintenance of silos where unMIRVed missiles are deployed. UnMIRVed missile silos (e.g., MM–II silos, etc.) could not be upgraded or deepened in any detectable way. Both sides could, of course, upgrade silos where MIRVed missiles were deployed.

In the case of the U.S., this means that we would have to forego the planned hardening of 300 MM–II silos. In the Soviet case, the impact could be more severe. The Soviets would have to forego hardening of up to 730 SS–11 silos, as well as their MLBM silos. This may overstate the impact since the Soviets have been upgrading their 200 older SS–11 silos in the Far East to take the SS–11 MOD–3 and this work could be completed. Also, the impact of this constraint will depend on the duration of the agreement. Nevertheless, this provision will undoubtedly cause problems with the Soviets.

In Sum

—We can accept deployment of both the SS–19 and 17 if the Soviets will confine this to designated areas and not make changes elsewhere.

—We should still try to ban the SS–X–18.

—Both sides could have the same number of ICBM MIRV fields, (i.e. 3).

  1. Source: National Archives, Nixon Presidential Materials, NSC Files, Box 891, SALT, SALT TWO–I–Geneva, January 1974. Secret; Sensitive; Exclusively Eyes Only. Sent for urgent information.
  2. See Document 44.