THE AGONY OF THE MX is over, for the moment at any rate. On Thursday, the House voted to release $1.5 billion for the purchase of 21 additional missiles after months of debate and acrimony. In the end, it was the president's argument of last resort that carried the day: How could his negotiators face the Soviets in Geneva without a bristling inventory of strategic hardware to bargain away?

But at what, precisely, is the Pentagon going to point these weapons, these "prompt hard-target killers," as defense officials call them? How do they fit into the Single Integrated Operational Plan -- the top secret U.S. war plan that establishes targets for the roughly 10,000 warheads in the U.S. arsenal of strategic nuclear weapons?

Astonishly enough, Congress does not know the answer to these questions. The cost, performance characteristics and basing mode for the MX have all been argued at exhausting lengths, but what we actually plan to do with the missiles in the event of war is taken entirely on faith.

There is no equivalent in Congress of the Intelligence Oversight Committees for review of American war plans contained in the SIOP, and existing committees do not receive thorough, regular briefings on it.

Yet the scientific community is now in agreement that a climactic catastrophe -- a "nuclear winter" caused by the smoke from burning cities -- could result from carrying out these plans.

Representatives and senators I have talked to seem vaguely sure somebody must be briefed on these matters. One House member vigorously doubted whether congressmen can be trusted with secrets of such magnitude -- as if a plan to obliterate the Soviet industrial infrastructure with several thousand nuclear warheads in the final stages of a nuclear war might be expected, unless revealed by some garrulous congressman, to come as a surprise.

This confidence that somebody has these matters well in hand is not limited to Congress. The whole national security establishment, from the president on down, seems to share it. But when you get particular, and look around for the people in charge, you find that no body, no agency or committee, no appointed or even self-appointed group in the White House, the Pentagon, or the Congress, has been asked to question the SIOP and its implications for the planet in rigorous detail on a regular and continuing basis.

Asked which members of Congress received briefings on the SIOP, the staff of the Joint Chiefs spent a week considering how to put it, and finally delivered the following written answer: "Those individuals/committees in Congress with a strict need to know are briefed on the overall policy and approach for the planning and employment of nuclear forces, when appropriate."

This awkward statement tells us three things: First, there is no briefing of committees as a whole on SIOP.

Second, such briefings as occur deal only with "overall policy and approach" -- windy strategizing -- not details about what we plan to hit, when in the unfolding moments of a war we would hit it and what effect this would have on the earth's atmosphere, climate and agriculture. Third, although the SIOP is revised about twice a year, there is no regularly scheduled briefing on the revisions.

In practice, according to congressional staffers, the chairmen of the House and Senate Armed Services and Appropriations committees are entitled to an annual briefing on the SIOP, and get it when they ask. But can an hour or so a year provide more than the most generalized notion of the nation's vastly complex nuclear war plans and targeting schemes? The answer seems self- evident.

Does this matter? Where nuclear weapons are concerned it is often hard to convince people that anything matters any longer. They know the weapons are too dangerous to use, we have too many, they cost a mint and don't make us safe, we can "destroy our world" umpteen times over, etc., etc.

The problem of getting a meaningful congressional reaction to the threat of nuclear winter is a case in point.

When the first nuclear winter study was published in December 1983, the public response from the White House and the Pentagon was silence. A group of distinguished scientists had said, in effect, that execution of our own war plans -- quite apart from anything the Russians might do in response -- might threaten the planet we live on. But the people in charge of national security said nothing.

Their silence had a simple explanation: They didn't know the truth or falsity of claims that burning cities, ignited by nuclear war, could pump enough dark, oily, sooty smoke into the upper atmosphere to envelope the Northern Hemisphere wit an opaque shroud for a couple of months or more, blocking out the sun and radically cooling the land by as much as 40 degrees Centigrade.

They didn't know because nobody involved in any stage of the war planning process had ever analyzed that possiblity. Nobody had ever thought to add up all the smoke from burning just about every city in Russia on the same day, or to ask where the smoke would go or to wonder what it would do.

If the authors of the original nuclear winter study are right, we crossed a kind of fatal frontier back in 1954 or 1955, when the United States first acquired enough warheads to trigger a climactic catastrophe, and made plans to use them. Nobody would have been more honestly amazed by the awful result than the planners themselves.

The situation has not changed. The current SIOP includes thousands of targets in Soviet urban areas, according to many sources. These targets are all "military" in one sense or another, but they are surrounded by combustible cities all the same. Most U.S. warheads aimed at these targets are scheduled to detonate at the "optimum height of burst" -- that is, in the air above the target -- to maximize the spread of blast and thermal pulse.

The thermal pulse of a nuclear weapon is a flash of heat so intense that just about everything of organic origin over an area of many square miles bursts spontaneously into flame. Air bursts are not incidental to the war plan. Along with blast, they are the plan.

Execution of the "major nuclear options" in the SIOP would mean hundreds of pillars of smoke from firestorms in Soviet cities. And the message of the nuclear winter scientists is unmistakeable: For our own safety, we ought to take Soviet cities out of the targeting plan. But the targeting fraternity knows that if we delete the cities, there is no plan left.

Defense Secretary Caspar Weinberger's own 17-page report on nuclear winter, released on March 1, conceded the problem was a real one, but insisted that no change in American war plans was called for. "The most basic elements of our policy remain sound," the report said. " . . . The United States must maintain a strong deterrent capability" -- that is, the capacity to inflict a devastating society-threatening nuclear attack on the Soviet Union.

This is deterrence in its purest form.

However, an attack that avoided destroying cities on a planet-threatening scale would still be devastating enough to deter any sane national leader. No American official ever argued in the past that we needed a planet-threatening "doomsday machine" to deter the Russians from attack. Do we need one now, simply because we have built one inadvertently?

The point to remember here is that American war plans -- and no doubt Russian plans as well -- call for attacks in the event of war on a scale that threatens disaster. The plans could be changed with relative ease, and the danger avoided, by reducing the number of targets in Soviet urban areas, and by switching from air bursts to ground bursts for those that remain.

The Joint Strategic Target Planning Staff at Offut Air Force Base near Omaha, Neb., which draws up the SIOP, can't elect to make a change of this sort on its own authority. The Pentagon has shown no sign of thinking such a step is necessary. Congress is completely excluded from the war-planning process. Thus we find ourselves in a curious situation -- threatening ourselves with our own war plans, but incapable of deciding to do anything about it.

Until 1960, plans for the use of nuclear weapons were adopted by each of the commands with control over nuclear weapons -- the Strategic Air Command at Offut Air Force Base; the Navy; and the Atlantic, Pacific and European Commands. In theory, their separate plans were coordinated, but in fact they were drawn independently and then held in the tightest secrecy.

Gen. Curtis LeMay, commander of SAC from 1948 until 1957, was notorious in military circles for refusing to tell anybody what he intended to do in the event of war, or even when he intended to do it. There is a good deal of evidence that LeMay had privately concluded that the United States ought to "preempt" -- that is, strike first -- if war looked probable. But, at heart, LeMay's strategic approach was simplicity itself -- hit 'em with everything we had. By 1957 that was plenty.

The planning anarchy began to change in 1960 when President Eisenhower was persuaded it made no sense to let each of the services plan to win the war all by themselves. To end the problem of overlap and duplication in targeting, he created the Joint Strategic Target Planning Staff (JSTPS) at Offut and asked it to draw up the first SIOP, which was approved that same December.

It called for a single, albeit coordinated, all-out attack -- everything in "one flush." Overkill was the order of the day. According to George Rathjens, now a professor at MIT, one Soviet city the size of Hiroshima was targeted with four nuclear weapons with a total yield of seven megatons -- more than 500 times the explosive force which destroyed Hiroshima.

In drawing up their plans, the targeters considered only the blast effects of nuclear weapons and worked within exaggerated guidelines calling for probabilities of kill high enough to justify SAC demands for more weaponry. The first SIOP also made little distinction between the Soviet Union proper and targets in Eastern Europe and China.

SAC's SIOP was rewritten after President Kennedy took office in 1961. Following guidelines laid down by Secretary of Defense Robert McNamara, planners set up four "options" defined by the types of target included in each. Moscow was "withheld" from initial attack plans, and Eastern Europe and China were no longer automatically included.

But for 10 years following approval of the second SIOP, the target planners at Offut were left in a kind of official limbo. New weapons entered the American strategic arsenal at a rapid rate throughout the 1960s, but no one told the JSTPS what to do with them. According to the physicist Richard Garwin, an Air Force adviser during that period, the target planners were reduced to reading the public speeches of the president and defense secretary for guidelines. "Hell," said a JSTPS officer in 1970, when MIRVed missiles aboard submarines rapidly increased the number of strategic warheads available for inclusion in the SIOP, "what are we going to do with all of those?"

The answer was to increase the number of Soviet targets on the National Strategic Target List, largely by refining and dividing existing targets. As the number of weapons and targets continued to grow, the "low" option in the SIOP -- the minimal strike planned in advance, which was aimed at Soviet nuclear forces -- grew right along with it. As late as 1974, the low option -- what we planned to do instead of an all-out attack -- called for hitting the Soviet Union with about 2,500 nuclear warheads.

In January 1974, President Nixon signed National Security Decision Memorandum (NSDM) 242, a five- page document adopting a new targeting strategy -- called the "Schlesinger doctrine" after then- Secretary of Defense James Schlesinger, although Henry Kissinger was its true source. In essence, NSDM 242 recognized that accurate MIRVs made it possible to target an enemy's strategic weapons and vice versa. This demanded a new approach that would suggest American readiness to make a limited response to a limited attack.

There has been endless public argument over the wisdom of this change in doctrine, but one thing is certain: it ended the isolation of the JSTPS, and provided a renewed sense of direction for the war planners.

President Carter approved a second strategy document in August 1980, Presidential Directive 59, followed in October 1981 by Reagan's signature on National Security Decision Directive 13. Each of these documents, none more than a few pages long, was forwarded to the Office of the Secretary of Defense where it was used in the preparation of a Nuclear Weapons Employment Policy (NUWEP) that went to the Joint Chiefs of Staff for further elaboration into a detailed guidance document intended for use by the JSTPS in writing a new nuclear war plan.

For obvious security reasons, not much has been published officially about the immensely complicated war-planning process. Many different types of weapons are involved: single and multi-warhead land-based missiles (the most accurate weapons in the U.S. strategic inventory), multi-warhead but not so accurate sea-launched missiles (SLBMs), and the various missiles and gravity bombs carried by manned bombers. Some are quick to reach their targets, some slow. Gravity bombs now have the largest explosive yield, SLBMs the smallest.

Executing the first SIOP in 1961 could have killed more than 400 million people. By 1977, the casualty estimate was down to 113 million. Part of the reason for the drop was Soviet civil defense measures, but most of it was the result of a change in approach -- a kind of tentative backing away from Armageddon encouraged by improvement in the accuracy of reentry vehicles.

But this does not mean that U.S. targeters plan to spare Soviet cities. Far from it. The current SIOP includes "limited nuclear options" of a few dozen or a few score warheads, the sort of attacks that might occur in the early stages of a nuclear war. But the major nuclear options involved in all-out attack would focus on so-called "recovery targets" -- things such as factories or railroad yards, and institutions such as the Soviet economic planning agency Gosplan or the KGB.

These targets are mostly in cities. The SIOP does not target population "per se," as the Pentagon never tires of saying, but it does target things surrounded by people in cities. According to a new book soon to be published by William Arkin and Richard Fieldhouse, "Nuclear Battlefields," the National Strategic Target List includes about 2,300 targets of an economic-industrial nature in the Soviet Union -- specific plants, power stations and the like -- which are covered by roughly 1,500 "aimpoints."

A 1978 study of Soviet civil defense planning by the Arms Control and Disarmament Agency (ACDA) suggested that a U.S. retaliatory strike during a major nuclear war would target 80 per cent of all Soviet cities over 25,000 in population. Moscow would be hit with up to 60 warheads, Leningrad with 40 or more, and the next eight largest cities in the Soviet Union with an average of 13 each.

The strategists' reason for executing attacks on this scale is simple: there would be no conventional "victor" in a nuclear war, neither side would be in a position to occupy the other, and once the shooting stopped both sides would be left willy-nilly to recover as they could. The first to recover an ability to wage general nuclear war would be in a position to dictate to the other, or so it is feared by the war planners.

The things that make a nation strong, and would allow it to recover, are mostly to be found in cities. As a result, the SIOP calls for society-crippling attacks on thousands of targets in Soviet cities.

This is a bleak fact, made even bleaker by the nearly universal belief among American military men and civilian defense officials that any nuclear war would tend to go the limit. The Russians go a step further. They say explicitly that they do not believe any nuclear war can be limited, and would respond accordingly to a limited attack.

Both sides have been investing heavily in communications infrastructure in recent years, thereby ensuring that the last aspect of modern society to collapse in a nuclear war would be the ability to carry on the war to the end. Thus, we might summarize the history of the last 40 years in the following way: first we invented a weapon of city-destroying potential, then we built thousands of them, found ways to deliver the first salvo in minutes rather than hours, skimped on conventional weaponry that might have let us put off using nuclear weapons and trumpeted our plans to annihilate our opponent, all but insisting he do the same.

Now we are told by many sober and reputable scientists that what we have come to think of as war may be more than the planet can handle. Skeptics have been invited to do their worst, but have been unable to fault the science. Many unknowns remain, of course, but no one, now and probably for years to come, can say with real assurance that carrying out our own war plans will not threaten human survival. Since this is the case, should we stick with the plans? Should we carry them out in the event of war?

Who is to answer these questions? With the arguable exception of Jimmy Carter, no American president has ever acquired more than passing knowledge of how we planned to use nuclear weapons in war. Eisenhower got an hour's briefing on the first SIOP, Kennedy roughly the same on the second. Johnson was impatient of the subject and Nixon no better.

In his volume of memoirs, "On Watch," Elmo Zumwalt describes a meeting of the National Security Council in January 1974, shortly after Nixon signed NSDM 242. It was clear, he says, that the president had no idea what his new strategy meant. Carter took a close interest in strategic matters and played a role in command post exercises to familiarize himself with authorization procedures and the like, but he remains an exception.

This should not be surprising: the SIOP is essentially a five-foot shelf of computerized data -- technical, abstract, highly particular, and dull. Even the category 10 SIOP briefing book used at the White House level is a couple of inches thick. No President has the time or expert knowledge to pass judgement on a plan so complex. He must depend on his advisers. Their record is not good.

Responsibility for drawing up war plans has been passed down the chain of command to the field grade officers who write the SIOP, men of great technical competence and devotion to duty, but no authority to decide where Mother Nature has drawn the line past which we must not go.

Procedures for oversight and review -- such as they are -- stick to the narrow military goals common to war plans. No body has ever been established to consider the full implications of carrying out our own war plans, and Weinberger's report to the Congress -- stressing deterrence pure and simple -- makes it clear that he, for one, does not see the need for any such regular and continuing review.

This leaves Congress. It will be argued that the SIOP is too complex for congressional oversight, that it is too secret to describe even in executive session, that authority for war planning resides in the president as commander-in-chief, and that it is unnecessary.

Similar arguments were made in opposition to Congressional oversight of the intelligence community. But the proper work of Congressional overseers would not be to approve the SIOP, much less to write it. Their job would be both simpler and more difficult -- to insist on knowing what is in the SIOP, to consider the full implications of carrying it out, and to satisfy themselves that U.S. war plans do not threaten our own security. They would serve, in effect, as a kind of circuit breaker. A prolonged failure to convince a Congressional War Plans Oversight Committee that the plans made sense would suggest something had gone seriously wrong.

Thinking about nuclear war is no fun. No one wants to look it in the eye. There is a temptation -- as strong among officials as it is for the rest of us -- to grant that a major nuclear war would be a terrible event and let it go at that. Congress has been content to fund the hardware, and let others decide what to do with it. Thus the buck has gone around, stopping nowhere. Almost absent-mindedly, we have planned a disaster. Perhaps now would be a good time to ask, with more energy than has been our custom, if our plans represent something we really want to do.