MaRV: keeping a new nuclear genie in the bottle

Public discussion of the deployment of United States Pershing II missiles in West Germany - scheduled to start in a few months as part of the just-reaffirmed NATO response to Soviet deployment of SS-20 intermediate-range rockets - has focused almost exclusively on relative quantities. But Pershing II has qualitative features which may be far more important than their quantity; features which not only can be, but should be, ''bargained away.''

There will, after all, only be a tiny quantity of Pershing II launchers: 108 of them. And the missile itself will carry only a single warhead in the Hiroshima range. (By contrast, the 300-plus Soviet SS-20s carry triple warheads each around 10 times as powerful; and both countries have long-range ICBM warheads hundreds of times as powerful.) While the Pershings will be less than 10 minutes from Soviet targets, roughly the same is also true of hundreds of submarine-launched missiles already deployed. So why all the fuss?

The key problem appears to be the Pershing's ''Maneuvering Reentry Vehicle,'' or ''MaRV,'' now undergoing tests. This allows the warhead to make terminal course corrections - by comparing internally stored data on the terrain with radar readings - and land within an expected 20 to 40 meters of its target from over a thousand miles away. This is about one-tenth the best miss-distance achieved by ordinary missiles, and it is the first time such a capability will have been given to any ballistic missile (new US cruise missiles have a similar capacity, but they are hours away from their targets).

The significance of MaRV can be seen in another measurement, called ''lethality'' or simply ''k,'' which assesses a warhead's ability to destroy a pinpoint target, like an underground command bunker, communications facility, or hardened missile silo. ''K'' is calculated on a formula which combines explosive power and accuracy (technically, the two-thirds power of the yield, in megatons, divided by the square of the ''circular error probability'' in nautical miles), and the best current ICBM warheads have a ''k'' around 100; most are considerably lower.

Pershing II's phenomenal accuracy, however, promises to give it a ''k'' somewhere between 100 and 600 per warhead. At the upper end of that range, the Pershing IIs would possess well over one-half the combined ''k'' of US existing strategic missile forces, and for the first time they would place within 10 minutes of the Soviet Union something which submarine-based missiles do not have: the capacity to destroy hardened targets, even if only 108 of them.

Though Pershing II, ironically, results from Russia's attempt to intimidate Western Europe with the SS-20, Soviet alarm over its characteristics seems genuine and understandable. MaRV is fundamentally undesirable. Not only would its deployment on Pershing II greatly heighten tension, it would let the ''MaRV'' genie out of the bottle, much as the ''MIRV'' genie was let out a dozen years ago to our own subsequent chagrin.

United States deployment will almost certainly be followed by Soviet testing and deployment. Sub-launched missiles are the likely candidate, and within a decade mutually MaRVed forces of such missiles will probably represent a brand-new ability to knock out essentially all hardened targets with only eight to 15 minutes' warning. Land-based missiles will have to be given up, or made highly mobile (and probably uncountable), and great decentralization of launching authority will almost certainly be needed to compensate for the extreme vulnerability of central command and communication facilities. ''Launch-on-warning'' policies may well be forced on both sides.

All this is clearly neither desirable nor remotely necessary. MaRV is an unneeded genie that should stay in the bottle. The place to start is by bargaining away this specific qualitative feature of Pershing II in return for Soviet concessions on SS-20s; this should be accompanied by a moratorium, and then a formal agreement, prohibiting all MaRV testing by both sides. NATO would still have the important political symbolism of some unmarved Pershing IIs (no more threatening to the Soviets than present sub-based missiles), plus a reduction - and it should be substantial - in SS-20s. Just possibly, a gravely destabilizing ''MaRV race'' can still be avoided.