War is politics continued by other means. So the Prussian military theorist Carl von Clausewitz said in the 19th century, in a definition that is perhaps less cynical than it sounds. He was challenging the notion that combat is purely a matter of violence, its outcome settled by whichever army had superior firepower. Yes, war involved an art (tactics) and a science (strategy). But there were larger forces in play.

Clausewitz had seen battle in the Napoleonic wars and did not overlook the brutal realities. But his hard-won realism included an understanding that the deeds of generals and diplomats were tightly linked. They influenced one another, each part of a much bigger process than either could fully master. (We owe to Clausewitz another memorable phrase, much in mind of late: "the fog of war.")

There was some debate during the Cold War over whether the old general's thoughts on strategy and tactics still applied in an age of mutually assured destruction. But we get a telling reminder of the durability of his deepest insights from The Atomic Bazaar: The Rise of the Nuclear Poor by William Langewiesche, a former national correspondent for The Atlantic Monthly.

Langewiesche's book originally appeared as articles in the magazine. But he has avoided the all-too-common tendency of journalists to bloat their prose when sandwiching it between hard covers. The Atomic Bazaar is a lean and cogent report on the emergence of a new "nuclear club" among the world's poorer countries. Alarming without indulging in sensationalism, it has the feel of a book that will prove prophetic.

The situation Langewiesche describes is shot through with bitter ironies. For decades, the United States and a few other nations accumulated nuclear stockpiles that (thank heaven) were never used. The weapons, and their delivery systems, were costly to build and maintain. The governments with such weapons tended to be stable and relatively prosperous (the one exception Mao's China).

Over time, though, the arsenals came to be more of a drain on each country's resources than a reflection of power. That created an incentive to de-escalate the arms race. And members of the "nuclear club" had every reason to want to keep it exclusive. International non-proliferation efforts offered to share technology for the peaceful use of nuclear energy with countries that agreed not to launch weapons programs.

So much for the old arrangement. The stability of the Cold War was a lot more precarious than now remembered. But it had its logic - much of which has been stood on its head.

Once the technical knowledge required to build atomic weapons was secret and scarce, and the logistics of starting a program to build them prohibitively difficult. Now the information is readily available. And for leaders of impoverished countries, the ability to prepare for nuclear warfare is, quite simply, a matter of common-sense politics pursued by other means. Rich countries want to restrict the availability of nuclear arms - which means that acquiring them is a sure way to get some respect on the international stage. Building one gives you an enormously valuable bargaining chip.

It also benefits domestic morale, especially in regimes that otherwise would not have much going for them. The capacity to menace neighbors with an atomic weapon is impressive in ways that the right to habeas corpus is not.

Then again, if you live next door to a country that is able to make such threats, what else is there to do but launch your own program? Likewise, all the incentives for an impoverished nation to acquire nuclear capability apply to "non-state actors" (not a term we used much during the Cold War) such as al-Qaida.

There is plenty to worry about here, but the value of Langewiesche's book is that it offers more than scenarios for concern. A case in point is the chapter called "Nukes Without Nations." It offers a rich account of efforts to monitor and disrupt the emergence of an international black market in the materials necessary to prepare a nuclear weapon, whether explosive (capable of vaporizing people) or dirty ("merely" poisoning them).

The news is neither completely terrifying nor even slightly comforting. A visit to nuclear facilities in the former Soviet Union, for example, shows that they are under serious if by no means foolproof security. He also hears from an American official about one such place "where informal parking lots have sprung up outside holes in the fence, because workers prefer not to bother with the gate."

Resources are allocated to watch for radioactive material at ports and border crossings. But someone with horses and a feel for the backcountry could probably manage.

"A U.S. government that could somehow find a way to lay traplines in [the] slums [of Istanbul, Karachi, and other cities where bombs could conceivably be built]," notes Langewiesche, "would have a better chance of stopping a terrorist attack than any port-inspection program, bureaucratic reshuffling, or military maneuvering can provide."

The danger of atomic weapons being created by non-state actors is, as yet, only potential.

So Langewiesche moves on to give an account of the career of Abdul Qadeer Khan, the mastermind of Pakistan's bomb. He is also the pioneer, if that is the word for it, of nuclear proliferation among other poor countries, for Khan's efforts established the international network needed to bring all the elements together.

To build the bomb, Khan had to engage in a certain amount of spying and dissimulation. But not as much as you might think.

"The manufacturers who sold to Kahn," Langewiesche reports, "like the European professors who signed on as his consultants, tended to be willingly naive and greedy. Those who were confronted by Western authorities invariably claimed to believe they were helping an impoverished country to pursue peaceful research."

While Khan is a national hero in Pakistan, his still-grander title is as "father of the Islamic bomb" (giving his country parity with India, with its presumably Hindu weapon). But his real legacy is more cosmopolitan still. Khan has admitted to sharing information with Iran, Libya and North Korea, and the network he created will doubtless outlive him. As Langewiesche puts it: "By its very nature - loose, unstructured, technically specialized, determinedly amoral - it is both resilient and mutable and can resume its activities when the opportunity arises, as inevitably it will."

The author's attitude toward all of this is not panic (readers may respond differently) but a steely realism that brings to mind Clausewitz. You can't build an international political perspective around the unpredictable role of non-state actors, terror-prone though they may be.

"It seems entirely possible," he writes, "that terrorist attacks can be thwarted - though this would require nimble government action - but no amount of maneuvering will keep determined nations from developing nuclear arsenals. North Korea, Iran, perhaps Turkey, Saudi Arabia, Brazil. Now and then a country may be persuaded to abandon its nuclear program, but in the long run, globally, such programs will proceed."

Understanding that means facing "a new reality in which limited nuclear wars are possible, and the use of a few devices, though locally devastating, will not necessarily blossom into a global exchange." That's the bad news. There's not really much good news, except that "possibility" does not mean "inevitability."

"The evidence so far is that even the poorest or most ideological countries are subject to the conventional logic of deterrence," as Langewiesche observes, "and will hesitate to use their weapons because of the certainty of a crushing response - since they, too, have cities and infrastructures that they will lose."

So proliferation, while worrying, "does not meet the category of threat that can justify pre-emptive wars." It won't if strategy and tactics are reality-based. But remember, sometimes war is the apocalypse continued by other means.