Foreign Relations of the United States, 1961–1963, Volume VII, Arms Control and Disarmament

118. Memorandum From the Deputy Assistant Secretary of Defense for International Security Affairs (McNaughton) to Secretary of Defense McNamara⁰

Washington, January 12, 1962.

SUBJECT

• Atmospheric test ban

Mr. Gilpatric asked Dr. Johnson and me to look into the question of an atmospheric-only test ban. He asked Dr. Johnson to study the question of how ready-to-test we could keep ourselves. He asked me to put that question in its broader context—i.e., with other factors which bear on a decision whether or not to resume atmospheric testing. Dr. Johnson ultimately joined forces with Dr. Seaborg, Dr. Wiesner and General Starbird to produce a report on his question for Mr. McGeorge Bundy on January 10 (you have a copy)¹ His results are summarized in my paragraphs 3(a)(iii), 3(b) and 3(c) below. This memorandum represents my results (I have deliberately oversimplified the issues, hopefully without distorting them in any relevant way).

1. Why NOT test in the atmosphere? The primary reasons for stopping atmospheric testing are (a) to dampen the arms race (especially by preventing the development of a reliable Soviet disarming or AICBM capability) and (b) to make a net propaganda-prestige gain. (c) The avoidance of harmful fallout, except as it bears on the propaganda battle, is not as important as the other reasons (see draft report of Federal Radiation Council).²

2. Must the U.S. test in the atmosphere even if the Soviets do not? We probably do need to test if we wish to maintain, as long as possible, nuclear superiority. (“Superiority” will probably still have some meaning—in fighting war, and in deterring Soviet lower-level adventures by fanning fear of a possible U.S. first nuclear strike—when our 1962 test results reach stockpile.) But we probably do not need to test in the atmosphere if we intend to permit things to evolve into a nuclear standoff and to handle lower-level Soviet threats by conventional (or other non-nuclear) means.

(a) Deterrence.

(i) Against a Soviet first nuclear strike. The Soviets are, even after the 1961 series, quite unlikely to build weapons and defense systems sufficient to permit them to carry out an effective first, disarming strike on us without expecting to receive unacceptable retaliation. It is conceivable since the 1961 Soviet series may have produced “surprises,” but it is unlikely. Whether they would be likely to create such capabilities after their next series of tests is another question. (For dissent, see JCSM-880-61.)³

(ii) Against lesser Soviet adventures. The Soviets are now capable of building weapons and defense systems sufficient to convince them (and us) that they can, despite our best disarming and defensive efforts, deliver on us an unacceptable retaliatory blow. A decision by the U.S. not to test in the atmosphere will, by giving up the 18 months mentioned in paragraph 3(a)(ii) below, probably accelerate the demise of the credible U.S. first-nuclear-strike threat.

(b) If deterrence fails. Since a major purpose of the atmospheric ban is to prevent development of reliable AICBM capability (i.e., to keep each side in a “hostage” condition), it follows that the ban could contribute greatly to the losses in the event of a war. This effect will be offset to the extent that the ban succeeds in dampening the arms race and therefore reducing the destructiveness of the war.

3. What PRICES do we pay for an atmospheric ban?

(a) Time lost in the technology race.

(i) The first “time price” is not properly chargeable to the decision not to resume atmospheric testing because it has already been paid and cannot be retrieved. It is the 7-months’ jump the Soviets got on us when they surprised us with their 1961 series. (The time periods mentioned in this and ensuing paragraphs are meaningful, but figurative. Here, for example, it is clear that, with respect to tests of AICBM systems, the Soviets got perhaps an 18-month jump on us. We will not be ready for such tests in the 1962 series.)

(ii) The second “time price” must be paid only if there is a moratorium but whether or not the moratorium sticks. It is the equivalent of 18 months in the technology race. This is the time difference between April 1962, when we will be ready to go, and the fall of 1963, when the Soviets probably will be ready with an equally mature series.

(iii) The third “time price” will have to be paid only if the Soviets break the ban in a surprise move. This price can be said to be the equivalent of 2 or 3 months in the technology race. This figure is based on the judgment that the highest state of readiness that we can maintain over time would put our “canned” series roughly 2 or 3 months behind the Soviet surprise series.

(b) Quality. The quality of U.S. tests will be lower, if “canned,” than they would be if testing were going on regularly. This is because of the effect of the ban on morale. The best people will work on what can be done rather than on what cannot be done. The quality differential will increase as time passes.

(c) Money. After the initial capital investment (which might be $15 to $28 million, depending on whether the development of Christmas Island should be included), the cost per year of remaining on quick-reaction testing alert would be between $1 and $1.5 million a year. No estimate has been made of the net dollar effect on the military budget of a decision not to test in the atmosphere (no atmospheric testing but vigorous underground testing, different weapons systems, etc.). The effect would probably be a money saving—perhaps a substantial one.

4. Will the Soviets accept a proposal to ban only atmospheric tests? Probably not—at least not in good faith. Their military interests, as contrasted with ours, probably make them more reliant than we on atmospheric testing. It is unlikely that they would agree to a deal in which we could continue underground testing.

(a) If the Soviets refuse a U.S. offer of an atmospheric ban, it would probably produce a net propaganda gain for the U.S.

(b) If the Soviets “accept” the ban for two or three years and then find a pretext to break it, we will have been “taken” for 20 months or so in the vital technology race. The risk of losing these 20 months of time and the incalculable increment of test quality is the big risk facing the President!

5. What TACTICS should the President employ if he decides he wants the atmospheric ban?

(a) WHO should propose it? The ban should be proposed by the U.S. This gives the U.S. more control over its timing and form, and it gives us the major propaganda value. However, if the Soviets proposed it, it might be more likely to “stick.”

(b) What FORM should the ban take? The ban should be in the form of a treaty. This makes it somewhat harder to break (which, by the way, may work to our disadvantage if we want to break it). It may also provide a better vehicle for obtaining the participation of France, China, etc.

(c) What should be the TIMING? The President should make the offer not long before the date when the tests are scheduled to begin. Optimum timing would be a month before D-Day or, on the present schedule, by March 1. This permits preparations to go on. But it heads off the great movement of planes to the Pacific; it avoids a big item of psychological damage to our testing people; it allows time for negotiations; and it does not appear to be an ultimatum.

(d) How can we structure pressures to make the ban STICK? Here are some factors (in addition to the disopprobrium which would follow from the breaking of a treaty) which could be employed to help make the ban “stick”:

(i) U.S. readiness to launch a major series in the atmosphere (although this readiness, which would surely be reciprocated, makes it physically easy to break the ban and therefore makes the ban fragile). U.S. can, for an indefinite period, maintain readiness on a 3-months’ basis with the quality of the “canned” series being degraded somewhat by the passage of time.

(ii) Maintenance of U.S. military position (A) strong and invulnerable enough so that the Soviets will not believe that with testing they can achieve a first-strike capability, but (B) not so strong that they will believe that we with testing can achieve a first-strike capability. (Our present near-to-disarming capability, for example, probably put pressure on the Soviets to conduct the 1961 series; this problem cannot be avoided as long as we rely on a first-nuclear-strike threat, instead of lower level capability, to deter Soviet non-nuclear adventures.)

(iii) UN resolution supporting the ban.

(iv) Military sanctions.

(v) Economic sanctions, possibly including forfeitures.

(vi) Vigorous information program designed to emphasize factors favoring continuation of the ban.

6. Would the ban be better LATER? If the proposal to initiate an atmospheric ban does not now appear attractive, the proposal may nevertheless be attractive (a) after the U.S. has completed its mid-1962 series, or (b) after both the Soviets and the U.S. have completed their next series after that. That is, the proper thing for the U.S. to do may be to conduct its mid-1962 series and then, when that series has been completed, to offer an option to the Soviets: (a) An atmospheric ban dating from that moment, or (b) one dating from the conclusion of the follow-on series by both nations, perhaps with the dates of the follow-on series to be set by the two sides so that they coincide. This would not pick up the 7 months we lost when the Soviets jumped the gun in 1961, but it would avoid further time losses.

7. How does the atmospheric ban compare, in attractiveness, with the US-UK all-environments treaty? It is important to note that both permit us to be taken by surprise, and therefore we should object to both unless the partial “opening” of the USSR involved in the US-UK treaty is worth the risk. In more detail:

(a) The atmospheric-only ban is as good in one respect (in both cases atmospheric tests are “self inspectable”).

(b) It is as bad in one respect (it allows us to be taken by surprise).

(c) It is not good in one respect (it involves no “opening” of the Soviet Union).

(d) It is not as bad in one respect (it does not allow the Soviets to perform, clandestinely, tests which we deny ourselves).

(See McNaughton-Rostow letter of 3 January 1962 on this subject.)⁴

8. Conclusions.

(a) Advantages and disadvantages.

(i) It probably would be to our advantage (but not necessarily, though probably, to the Soviets’ disadvantage) if atmospheric tests could be stopped permanently now.

(ii) It would probably be to our advantage (and probably to the Soviets’ disadvantage) if we make the offer and they refused it.

(iii) It would probably be to our disadvantage (and to the Soviets’ advantage) if we made the offer and they “accepted” it, but the ban lasted only (say) six months to four years—long enough to cause us to sacrifice much or all of our present readiness-to-test advantage but not long enough to be worth, in terms of dampened arms race, that sacrifice.

(b) Structuring the deal. If there is to be an atmospheric ban, it must be structured to stick. It must be created and nursed along with that thought in mind.

John T. McNaughton

0. Source: Washington National Records Center, RG 330, OSD Files: FRC 66 A 3542, Atomic 400.112 12 Jan 62. Secret.
1. Not found.
2. Not found.
3. Not further identified.
4. A copy is in Department of State, S/P Files: Lot 69 D 121, Atomic Energy-Armaments 1962. For text, see the Supplement.