CHAPTER 7
SECRETS OF CHAKLALA CANTONMENT

We were compelled to show in May 1998 that we were not bluffing, and in May 2002, again, we were compelled to show that we do not bluff.

— Former President Pervez Musharraf, June 17, 2002,
quoted in a plaque on Khalid Kidwai's office wall

TO GET TO the headquarters of the blandly named Strategic Plans Division, the branch of the Pakistani government charged with keeping the country's growing arsenal of nuclear weapons out of the hands of al Qaeda and its militant Islamist sympathizers, you must drive down a rutted, debris-strewn road at the edge of the Islamabad airport.

Stray dogs and the homeless wander along a street lined with crumbling and collapsing houses, some of which are just piles of brick with corrugated metal roofs held down by heavy stones. Garbage piles up, uncollected, for months. In the distance you can see the haze of exhaust that seems to hover permanently over the nearest town, Rawalpindi.

Just past a small traffic circle, a tan stone gateway looms, manned by a lone, bored-looking guard loosely holding a rusting rifle. It marks the entry to Chaklala Cantonment, an old British garrison from the days when officers of the Raj escaped the heat of Delhi for the cooler hills on the approaches to Afghanistan. Pass under the archway, and the poverty and chaos of modern Pakistan disappear.

Chaklala is the well-tended home of the country's military and intelligence services, a reminder that in Pakistan the bulk of power rests with those two institutions, sometimes operating with brutal efficiency, but just as often undercutting each other. Not surprisingly, both the army and the Directorate for Inter-Services Intelligence, known as the ISI, have reserved society's best privileges for themselves. Inside the gates, they live in trim houses with well-tended lawns. Business is conducted in long, low office buildings that look like a single-level motel, with a bevy of well-pressed adjutants buzzing around.

About three-quarters of a mile down the road within the walls of the garrison and barely marked, lies the small compound for Strategic Plans, where Khalid Kidwai keeps the country's nuclear keys— and watches for any sign that Pakistan might, for the second time in recent history, be vulnerable to another major breach of its nuclear secrets, another attack from within.

The reality of the Second Nuclear Age is that, in the end, what happens or fails to happen in Khalid Kidwai's modest compound on the edge of the Islamabad airport is far more likely to save or lose an American city than are the billions of dollars we spend each year maintaining a nuclear arsenal that will never be used, or the thousands of lives and hundreds of billions of dollars we lost in Iraq. Yet Pakistan's ability to control its arsenal ranks among the least understood, least discussed, and in many ways the scariest nuclear challenges facing the next president—and, if we continue to be lucky, presidents for years to come. For while there are reasons to worry about loose nuclear material in Russia—a situation that remains worrisome, but less worrisome than it was a decade ago—Pakistan is the only nuclear state with a powerful militant insurgency in its midst, one that clearly has aims to take over the country, and desperately wants to acquire the Bomb. It does not help matters that the government has veered between a dictatorship that has supported both the United States and the Taliban and a newly elected democratic leadership known chiefly for its corruption and ineptitude.

Kidwai is fifty-eight, a compact man with an arch sense of humor, often hidden beneath a veil of caution, as if he were previewing each sentence to decide if it reveals too much. He was installed in his post, perhaps the most sensitive in Pakistan, by Musharraf himself, and for the nine years of Musharraf's rule, the two men remained close. Their worldviews were shaped by the Pakistan Army. They also share the army's demeanor—a preternatural calm meant to convey to Pakistanis and the outside world that everything is under control, no matter how bad things get. In 2007, both shed their uniforms but held on to their civilian posts-Musharraf as president, Kidwai as keeper of the country's nuclear arsenal. Musharraf's luck ran out, however, in August 2008, when he was forced to step down.

But Kidwai has held on to his job—a fact that some insiders in Washington consider far more consequential to American interests than who is sitting in the presidential palace. Kidwai oversees the entire security structure meant to keep Pakistan's nuclear weapons and fuel out of the hands of outsiders—Islamic militants, al Qaeda scientists, Indian saboteurs, even American Special Forces teams that the Pakistanis fear are perpetually bobbing just offshore, refining their plans to snatch Pakistan's weapons if a crisis erupts. (It is not an entirely unwarranted fear.)

In Washington, American officials know Kidwai because he was on the receiving end of one of the most tightly held secret projects with Pakistan since 9/11: a classified, nearly $100 million effort, financed by American taxpayers, to teach Pakistan how to lock down its nuclear- weapons. Even though American officials had only Kidwai's assurances about how the money would be spent—auditors were prohibited by the Pakistanis, who refused to let foreigners into their most sensitive nuclear sites—Bush determined it was worth the risk. The alternative was to do nothing to help secure an arsenal of upward of a hundred nuclear weapons, in the most volatile corner of the world. The program was one of the reasons that American officials insisted, during Pakistan's descent into chaos in late 2007, that they had no reason to doubt Pakistani claims that the nuclear infrastructure was secure.

Privately, though, they admitted they had plenty to lose sleep about. While there was little doubt the weapons themselves were relatively safe, there was plenty of reason to wonder about the security of Pakistan's laboratories, including the one still named for A. Q. Khan, years after the conniving metallurgist had been disgraced for his sale of Pakistan's nuclear secrets.

There was more—a threat whose severity President Bush and his top aides spent much of 2008 trying to assess. In the spring of that year, as Musharraf was losing his grip on power, an urgent new stream of reporting began coursing its way through the CIA, the Pentagon, and the White House. Al Qaeda and other militant groups were focusing anew on the Holy Grail that had eluded them before 9/11: stealing the secrets to the Pakistani bomb.

In the ‘90s, al Qaeda had fallen for at least one nuclear scam, buying up a box full of useless radioactive junk. Now there was evidence suggesting that al Qaeda and other terror groups were attempting to take a different road to the same destination. They were recruiting Pakistanis who had been trained in nuclear sciences and engineering abroad to try to figure out which ones might harbor sympathy for radical Islamic causes. According to an American intelligence report that was restricted to very senior officials in Bush's war cabinet, a few of those scientists appeared to be returning to Pakistan to seek jobs within the country's nuclear infrastructure.

By the spring of 2008, the entire top tier of the national security leadership in Washington had been briefed on the intelligence assessment. “I have two worries,” one senior official who had read all of the intelligence with care told me, as the reports were circulating. “One is [what happens] when they move the weapons. And the second is what I believe are steadfast efforts of different extremist groups to infiltrate the labs and put sleepers and so on in there.”

Quietly, American intelligence officials alerted their Pakistani counterparts, including Kidwai, about the threat. But like the warnings of airplane plots prior to 9/11, none of it seemed “actionable”: There were no specific names or places.

“This is all overblown rhetoric,” Kidwai told me soon after we settled into the big white leather chairs in his spacious office inside the garrison, decorated with models of the missiles that can carry nuclear payloads to India and beyond. It was a Saturday morning, and things were quiet at the Strategic Plans Division. He had time to pace me through the layers upon layers of protections that he and Musharraf had pieced together as Pakistan moved from nuclear pariah to the world's eighth nuclear power.

“Please grant to Pakistan that if we can make nuclear weapons and the delivery systems,” Kidwai said, gesturing to the models of the missiles and a photo of Pakistan's first nuclear test, a decade ago, “we can also make them safe. Our security systems are foolproof. This is what gives us the confidence that what happened before with Dr. A. Q. Khan—it could never happen again.”¹

“WHAT HAPPENED BEFORE” is the greatest known breach of nuclear security in the atomic age, a breach the Pakistanis calmly denied for years was happening at all.

Apart from the country's founder, Mohammed Ali Jinnah, who died in 1948, Pakistan may have no more revered hero than Khan, now seventy-two and in remission from prostate cancer. It was Khan—metallurgist, egoist, first-class self-promoter—who crowned himself “Father of the Pakistani Bomb,” a title to which others could rightfully lay equal claim. Most Pakistanis, however, do not care about the shameless self-promotion, or even about the brazen crimes Khan committed as he sold the fruits of Pakistan's nuclear program around the globe. In Islamabad, some taxi drivers still keep a picture of him dangling from their rearview mirrors.

For the roughly 160 million citizens of Pakistan, the popular narrative recalls how Khan, through a mix of stealth and savvy, figured out how to even the score with India and Western powers by turning a poor, fractious country into a nuclear player by the late 1980s, a decade before its nuclear test. Since that test, Pakistan has made the leap from international outlaw to an accepted, acknowledged nuclear power that the world must respect and fear. Even today, one widely read Urdu-language newspaper in Islamabad runs a daily feature about how many days Khan has been kept under house arrest, albeit in luxurious incarceration, and berates the government for failing to free him.

That protest may soon prove successful: In the summer of 2008, a Pakistani court relaxed some of the restrictions on Khan, though it ordered him to stop talking about his past. No doubt that gag order came as a relief to many Pakistani politicians and military officers, who are a lot more concerned about what Khan says than about what he does. They have reason to worry. Tired of his incarceration, embittered at Musharraf for pardoning him and then imprisoning him, perhaps emboldened by the sporadic protests suggesting he should have been elected Musharraf's successor, Khan had begun hinting that he was getting ready to talk. Specifically, he threatened to talk about who inside the country's power structure had been aware of what he was doing, and who was complicit. For example, he said in one conversation, there was a shipment of centrifuges—the giant machines that spin at supersonic speeds, enriching uranium into fuel—that he sent to North Korea when Musharraf was still chief of the army.

“It was a North Korean plane, and the army had complete knowledge about it and the equipment,” Khan said. Did that mean Musharraf, America's ally, knew? “It must have gone with his consent,” Khan responded, offering no evidence.²

Musharraf, of course, denied knowing anything, and in a lengthy interview in 2005 he told me that he had courageously taken on Khan. But whatever the truth, Khan brilliantly exploited his hero status, creating a culture of deference and secrecy that completely enveloped him and the bomb project into which Pakistan poured its sparse treasure. He was protected by a political and military elite that desperately wanted the bomb and were happy not to ask questions about his deal-making or demand to know how a Pakistani bureaucrat on a government salary could afford a house in Islamabad's best neighborhood and properties abroad. Their deliberate absence of curiosity gave Khan the opportunity he sought.

In the 1980s, when Pakistan already possessed a crude device but was still a decade away from testing it, Khan began to build a multinational network—stretching from Kuala Lumpur to Dubai and Capetown—that packaged and sold the nuclear technology he had accumulated throughout the years. Over time, this illicit project went far beyond selling the designs for uranium enrichment technology that Khan himself stole from Europe decades ago for Pakistan's own indigenous weapons program.

Khan was once at the center of almost every major nuclear flashpoint the United States faces today. It was Khan who sold Iran the uranium enrichment equipment and designs that put it within reach of a bomb and on a collision course with the West. It was Khan who sold the North Koreans those centrifuges, which they thought they needed to explore a new path to making weapons, after their first efforts—producing plutonium from an aging nuclear reactor-were shut down by the Clinton administration in the mid-1990s. And it was Khan who not only delivered the Libyans more than $100-million worth in centrifuges but threw in a bonus, wrapped inside the plastic bags from the dry cleaner near his house: the nearly complete blueprints for the bomb China set off in the mid-1960s. After a brief standoff in Tripoli between American officials and international inspectors over who should hold on to the confiscated designs, they were flown back on a special flight to Dulles Airport and stored in a secure underground room in the Energy Department, off the National Mall in Washington, as investigators tried to figure out who else had gotten access to the plans.

It turned out Khan had other designs as well. Deep in his computer systems he hid other, far more sophisticated blueprints. To the investigators who uncovered them long after Khan was already under arrest, the designs appeared to be for the Pakistani bomb. To this day, even while publicly celebrating the shattering of the Khan network as a rare coup for the intelligence community, investigators around the world are struggling to understand who else bought these far more complex plans, the recipe for the Pakistani arsenal. As I traveled the world exploring potential national security crises, the Pakistani bomb designs that got away elicited the longest silences, the most artful evasions, and, among those willing to talk, the most concern.

ASTOUNDINGLY, Khan's perfidy unfolded within miles of Kid-wai's headquarters. Khan Research Laboratories, where the power-hungry scientist built his empire and shipped out his equipment, is just down the road from Chaklala Garrison. From the highway you can't miss the laboratory's entrance: It is marked by a replica of a Pakistani missile, one of many government-financed monuments to the accomplishments of a man still considered a hero by most Pakistanis. On the other side of the garrison lies the airstrip Khan turned into his own personal FedEx hub, where a mix of charter planes and Pakistani Air Force cargo craft—provided to the country by the United States—were used to ship Pakistan's nuclear centrifuges and other equipment to clients who paid him hundreds of millions of dollars. (The exact amounts have never been made public.) To this day, the Pakistani military swears they had no idea what Khan was doing—an assertion American and British investigators have found difficult to believe.

Venture a few miles beyond the gates into the chaos of Rawalpindi, and there are constant reminders of why the world fears what could happen if Pakistan melts down. As al Qaeda and the Taliban have seen their chance to destabilize Pakistan, Rawalpindi has become an assassin's playground. Twice it was the site of efforts to blow up Musharraf's motorcade, a plot later tied to both al Qaeda and members of Pakistan's own military. By 2007, extremists had turned their sights on Benazir Bhutto. They attacked her and her followers on the day she returned to Pakistan to run against Musharraf. She survived. But it was only a matter of time, and in late December 2007, as she emerged from a decrepit park where she had just addressed thousands, they succeeded. The second attack, a mix of bullets and a suicide bombing, was as messy and chaotic as the country. It was a moment of raw violence that echoed Pakistan's history: The park Bhutto had spoken in was named for the country's first prime minister, Liaquat Ali Khan, who was assassinated fifty-six years before, just a few hundred yards away. (In one of those bizarre twists of history, the doctor who tried, unsuccessfully, to revive Bhutto at Rawalpindi's main hospital was the son of the doctor who tried, unsuccessfully, to revive Liaquat Khan.)

Those assassinations were products of different but linked conflicts that rip at Pakistan's core. Liaquat Khan was assassinated by a Pashtun separatist, in retaliation for the prime minister's insistence that Pakistan must extend its writ of control over the restive tribal lands along the border with Afghanistan. A century and a half ago, those same lands were a hotbed of opposition to the British. Today they have become the new sanctuary of al Qaeda, the Taliban, and a second generation of militants like Baitullah Mehsud, who has been accused by both the Americans and the Pakistanis of dispatching Bhutto's assassins. Three months after her death, the Pakistanis I met near the park believed—rightly or wrongly—that she was killed in part because she was an American puppet who talked of targeting those militants, and even of letting international investigators interview A. Q. Khan, something Musharraf had strictly forbidden. Even Musharraf, who was denying charges that his government had failed to protect his rival, couldn't bring himself to fake much regret about how she met her end.

“She was very unpopular with the military. Very unpopular,” he said a few weeks after her death. She was seen, he said, as an “alien” by Islamic leaders, flaunting the degree to which she was “a nonreli-gious person.” And she forgot the first rule of Pakistani politics: “Don't be seen as an extension of the United States. Now I am branded as an extension,” he said, “but not to the extent she was.”³

Whatever its motivation, Bhutto's assassination reinforced the sense in the West that Pakistan remains one anarchic turn away from spinning out of control. The summer before her death, there had been the bloody shoot-out with militants at the Red Mosque in Islamabad, the sudden appearance of suicide bombers in the heart of the capital, followed by the power struggles on the streets between Musharraf's forces and lawyers seeking the restoration of judges that Musharraf had dismissed for challenging his power. Taken together, all of these events raised fundamental questions few in Washington wanted to discuss publicly: Could a modern nuclear state lose control of its arsenal? Could another A. Q. Khan or, worse, Islamic fundamentalists gain access to nuclear designs or fuel? And if the worst case came to pass, was there anything we could do?

THE MAN WHO THINKS about these questions every day is Rolf Mowatt-Larssen, a former CIA officer who helped crack the Khan network and confidant of the former director of the agency, George Tenet.

These days Mowatt-Larssen has moved into the center of Washington—the National Mall, ground zero if trouble breaks out—working from an office within a secure vault in the basement of the Energy Department. The department is responsible for developing and maintaining the country's nuclear stockpile, and its intelligence division, which Mowatt-Larssen runs, has access to the government's best technical expertise on what it takes to make a bomb, and how to keep someone from getting what they need to build one.

In the corner of his office, where many people would keep a potted plant, Mowatt-Larssen keeps a centrifuge standing against the wall—one of Khan's, given to him after the Libyans surrendered their purchases from the Khan network. It is a reminder, one he sees every time he looks up from his computer, of how easy it is these days for shadowy networks to deal in the technology that was once the sole property of nations. Mowatt-Larssen is paid to design the worst-case scenarios about how quickly a country or a terror group could move from centrifuge to bomb—and then to test those assumptions against the fortunate reality that it's all harder than it looks.

But he also digs deeper, into the statements and philosophy of Islamic radicals, and sometimes comes out in some interesting places. I asked him about the assumption, taken as an article of faith among many that if al Qaeda ever bought or made a nuclear device, it would use it immediately.

“I don't think we know if that's true,” Mowatt-Larssen told me. “There would need to be some religious rationale first,” he argued, especially now that an alternative narrative has broken out in the Islamic world, one in which al Qaeda's tactics are creating a backlash that could be undermining its own goals.⁴

He described to me the struggle al Qaeda went through to devise the proper religious justification for obtaining and using a weapon that might kill the faithful along with the infidels. The high-water mark of that effort may have come in 2003, he noted, when al Qaeda obtained a fatwa from a sympathetic Saudi cleric, Nasir bin Hamid al-Fahd, who argued that while the use of nuclear or biological weapons by infidels would be unlawful, it was permissible by those defending the Muslim faith. Saudi authorities, as part of their effort to stamp out the al Qaeda forces that have tried so hard to oust the al Saud family, forced him to appear on television later that year to recant some of his rulings.

But Mowatt-Larssen is convinced that al Qaeda would have pressured a cleric to make such a ruling only if it thought it needed a religious justification for future actions. From his basement office, he monitors not only the movement of nuclear materials, he monitors the movement of Islamic opinion. An ideological war has broken out among Islamic militants over the question of whether Islamic law allows the kind of violence that made al Qaeda famous. There are some signs Ayman al-Zawahiri, bin Laden's second in command, may have held back on a chemical weapons attack a few years back because he feared the backlash.⁵ But no one is willing to count on Zawahiri's sense of restraint.

Instead, war planners in Washington focus on capabilities, and that is where Pakistan—and Kidwai's operation—become the central concern. Pakistan is the place where a fractious and corrupt government, an abundance of weapons, nuclear fuel and designs, and fatwas like al Fahd's all coexist within a few hundred square miles around Islamabad. And so, very quietly, because the backlash by the Pakistanis would be so severe, Mowatt-Larssen has tried to test the question of what happens if suddenly those weapons are in play.

That journey has led his intelligence operation inside the Energy Department, along with colleagues at other intelligence agencies, the Pentagon, and Special Forces commands, to conduct a series of what they benignly call “tabletop exercises” in which a chunk of nuclear material is suddenly assumed to be missing somewhere in Pakistan. Virtually every time, the scenarios end in a murky sea of ambiguity. Top officials in Islamabad and Washington can't get a clear picture of what happened, or of what's happening next, much less decide what to do about it. And by the time a clear picture does emerge, it is too late to do much except panic.

“The fear is that sometime in the next few years, someone is going to have to wake up the president and tell him that we think— we don't know, we think—that ten Pakistani weapons are missing, and we can only suspect who has them,” said David Rothkopf, a former Clinton administration official who wrote one of the leading studies of decision-making within the National Security Council. “But it doesn't play out like it does in the television shows. Not only can't you order a preemptive strike, you don't know where to strike. Or even what's really missing. It's the classic post-Cold War nightmare.”

Ashton B. Carter, another former Clinton administration official, has taken the scenario one step further for Harvard's Preventive Defense Project: He has assembled government officials to play out the scenario the “day after”—that is, the day after a weapon goes off in an American city.

“There's a 90-percent chance it will come out of the old Soviet arsenal or out of Pakistan's program,” says Carter. “And once you figure out which one it came out of, what do you do? Launch a nuclear strike against an ally for something the president of the country probably didn't know was missing? When's he's still got a closet full of nuclear missiles? Not going to happen. And so that leaves the only other option: Find a way to lock it down now.”

CARTER IS RIGHT: When the biggest threat looks more like loose nukes that escape Pakistan than launched nukes out of Russia, all the old tricks for avoiding Armageddon don't work. Our nuclear arsenal has become the Maginot Line of the age of terror: big, scary, and fundamentally useless as a deterrent. It was designed for a different age, for weapons that came streaking into the United States from silos in the Soviet Union, tracked every step of the way on a giant monitor deep in the mountains of Colorado. However, if a bomb gets out of Pakistan, it's unlikely to be on a missile, and it won't be showing up on that screen.

Every television producer worth an Emmy Award figured this out years ago, spawning the dozens of movies and series in which bad guys get a load of uranium or plutonium, and the only way to stop it is to send Jack Bauer into their basement. The drama hinges on whether the traffic is so bad on the Los Angeles highways that he'll arrive too late. By comparison, going around the world trying to fix every decrepit nuclear storage site, building double fences and training local police to use radiation detectors doesn't make for very exciting television.

Apparently, it turns out not to make for very exciting policy, either. Wherever I have traveled around the world—talking to American intelligence officers in Pakistan, military officials who patrol the Pacific, or the Homeland Security officials who spend their days thinking about attacks on the nation's capital—they all say the same thing: As the situation in Iraq worsened, the post-9/11 efforts to create a multilayered defense against a domestic WMD attack waned, even though Bush and his aides readily acknowledged that no terror group on earth could pose an existential threat to the United States unless they obtained a nuclear or biological weapon.

The United States spends a little more than a billion dollars a year locking down nuclear material outside the United States—not a small amount of money, but only a tenth of what we spend on missile defense, a technology that assumes the next attack on the United States or its allies will come streaking across the sky the old-fashioned way.

It is not that nothing is being done—there are new bureaucracies, chiefly the behemoth born of 9/11, the Department of Homeland Security. But the easy fixes are over, and the hard work of investing in a truly multilayered defense against nuclear and biological attacks lies ahead.

“Look, I don't think you are going to see someone drive a commercial airliner into buildings again. It could happen, but I think we've sealed that up pretty well,” said Adm. Timothy Keating, who headed the newly created “Northern Command,” the military command set up specifically to coordinate the military response to another 9/11. (He is now in charge of all U.S. forces in the Pacific.) What worries Keating is that the United States may once again be experiencing a failure of imagination, this time about a nuclear weapon flown into the country “on a private G-5 [Gulfstream V] aircraft, for example, or the ease of a biological attack. I worry about that, even if you have to say it is unlikely.”⁶

It is only one of the huge loopholes left in America's defenses. “It amazes me,” said John F. Lehman, a Republican and secretary of the navy under Ronald Reagan, who looked at this issue as a member of the 9/11 Commission. In December 2005, a year after the commission issued its findings, its members reconvened to issue a report card on how well the Bush administration and Congress did at sealing up the huge vulnerabilities discovered during the investigation. The report card was a sea of C's, D's, and F's, including a D for one of the most important recommendations: a “maximum effort” by the U.S. government to secure the world's nuclear weapons. The report concluded that “countering the greatest threat to America's security is still not the top national security priority of the president and Congress.”

I ran into Lehman in May 2008, when he was actively advising John McCain and asked if he had seen any improvement. “No,” he said. “It's gotten worse.” Our investment in real defense, he noted, defense against the single weapon that gets smuggled into a city in the United States or another city around the world, remained pitiful. Our strategy for dealing with Pakistan and its weapons was in disarray, to be generous. The result is that the United States has stumbled along with no real plan, just a series of largely disconnected, underfunded programs.

To his credit, Bush started one truly great innovation in this area. It was largely the invention of Robert Joseph, who in Bush's first term drove the White House to focus far more on “counterpro-liferation” (actively intervening to destroy or seize weapons) than on “nonproliferation” (keeping countries from going nuclear). Conservatives and liberals have been in an endless debate about which one is a smarter approach; the reality is that both are needed, simultaneously. No future president, Democrat or Republican, will ever have the luxury of picking one or the other.

Joseph's creation was called the Proliferation Security Initiative, and it involved informally signing up countries to intercept nuclear or biological shipments. Joseph deliberately steered clear of treaties or negotiated agreements “because they take ten years to happen,” he once told me. Instead, he focused on the laws each country already had on the books and pressing leaders to view those authorities more broadly. So far, more than seventy countries have signed on, some with more enthusiasm than others, and periodically members of the group hold exercises to board ships or intercept airplanes, demonstrating that they are serious about working together.

There are holes in the system: Pakistan refused to join, fearing it would be a target rather than a partner. South Korea refused to join, fearing it would anger the North. But enough countries did sign up that once a shipment of missiles to Syria or centrifuges to Iran are detected, there is an international procedure for stopping the ship or plane, especially one that stops somewhere for refueling.

That marks a huge advance. Until the Bush administration pressed countries to sign up, there was no effective way to halt such shipments. But by definition, the Proliferation Security Initiative is a backup measure that involves chasing down materials once they are en route to another country. It depends on superb intelligence and quick response. You cannot count on either.

That is why cutting off new supplies of nuclear fuel is so important, and fleetingly it seemed that Bush would be serious about getting something done. In 2004—right after the Khan network was broken up—he went to the National Defense University in Washington to lay out a seven-part plan to limit trade in nuclear material, calling for a halt on such shipments “to any state that does not already possess full-scale, functioning enrichment and reprocessing plants.” Of course, the language created a double-standard: Countries such as Iran would be prohibited from making nuclear fuel, even though they are signatories to the Nuclear Non-Proliferation Treaty. Any of America's friends who are already in the business— Japan or European countries, even nuclear renegades such as Pakistan—would be allowed to continue.

Some of Bush's closest allies began objecting. Canada and Australia, two countries that contributed critical forces in Afghanistan and Iraq, argued that they wanted to start their own enrichment operations to extract additional revenue from the uranium that they mined. Soon Bush simply stopped talking about his plan. It disappeared from the agenda at international meetings. Like Bush's speech about a Marshall Plan for Afghanistan, or a national effort to rebuild New Orleans, his main proposal to limit the amount of nuclear fuel produced in the world fizzled.

Then Bush did something that made things worse: He struck a deal with India that would allow the country, for the first time, to get civilian nuclear technology from the United States. For decades, such trade has been illegal because India never signed the Nuclear Non-Proliferation Treaty—one of only three countries that refused. (Pakistan and Israel also never signed, and North Korea quit the treaty in 2003.) American law made it illegal for any American firm to help India's civilian nuclear power program. For years the Indians had lobbied in Washington to get the ban lifted, arguing that the world's most populous democracy deserved better from the United States.

Eventually, that logic won out. In his second term, looking for a way to deepen the relationship with India—partly as a hedge against a rising China—Bush agreed to negotiate a deal that would allow American firms to export technology to India—worth billions— along with nuclear expertise. But the negotiations went terribly for the United States. Bush and his chief negotiator, Nicholas Burns, were never able to extract a promise from the Indians that in return for American assistance they would stop producing weapons-grade nuclear fuel and stop expanding their arsenal. Desperate for a deal, Bush signed anyway.

Of course, every kilogram of nuclear fuel that the United States sold the Indians to use in their power plants freed up some fuel to make more weapons. Pakistan, of course, vowed that if the Indians built more weapons, so would they. Bush, in short, accelerated the arms race in South Asia.

“This took stupid to new levels,” a senior American military official in Islamabad said to me in the spring of 2008. “We're going into the Paks every day and warning, ‘Look, you have to lock up your weapons and all your fuel, because the more there is, the better the chance that one day you are going to wake up and discover Osama's got some of it.’ And they say, ‘That's your problem. You're helping the Indians, and wasn't it Bush who said we are America's great ally against terrorism?’

“Who are we kidding?” the senior military official asked. “Musharraf? Ourselves?”

The result is that Kidwai sits in his office overseeing two separate missions. One is to plan Pakistan's nuclear future, as it keeps building more and better weapons. The other is to convince skeptical visitors that what happened a few years ago could never happen again—even if Pakistan goes up in flames.

KIDWAI'S PROBLEM is that every time he tries to convince the world that in Pakistan everything is locked down, people familiar with the country's history have the same reaction: We've heard it all before.

When the Times published its first lengthy investigative piece about Khan's activities in January 2004, a Pakistani government spokesman reflexively denounced the article as “a pack of lies.” A month later, Khan was forced by Musharraf to “confess,” though he never said exactly what he was confessing. He was never charged, never tried, and never convicted. He was simply pardoned and put under house arrest—the best way to keep the details of how he raided Pakistan's nuclear secrets out of the press. When the CIA or international investigators had questions for him, they had to route them through the Pakistani government. The answers they got back, one of the top investigators said to me one day, “are the answers the Pakistanis want us to hear.”

Today Kidwai argues that the Khan era is over and the West just does not recognize the progress the Strategic Plans Division has made. Though he will not go into details, he nods in silent agreement when asked if American experts are correct in assuming that Pakistan stores its missiles separately from its warheads, and its warheads separately from their nuclear triggers, making it more difficult for terrorists to acquire all the components they need.

He keeps a PowerPoint presentation at the ready in his conference room to show how 2,000 scientists with “critical knowledge” of nuclear technology are constantly monitored, vetted, and subjected to psychological profiling. In recent years the monitoring has extended to retired members of the program to make sure that, in addition to their pensions, they don't collect a little something extra from a richly endowed Islamic militant group in return for their expertise.

Both in public and in private, White House officials say they are impressed with what the Pakistanis have done, and they hint with knowing smiles that they know more about the status of Pakistan's weapons than they can say. Nonetheless, dig a little deeper and one quickly discovers doubts about Kidwai's confidence that “it could never happen again.”

Put plainly, most senior American officials who track nuclear issues think Kidwai is under such pressure to put the best possible face on Pakistan's nuclear security that he glosses over the most potent threats.

As the 2008 intelligence reports that shot through the top of the administration indicate, analysts worry most about Pakistan's nuclear laboratories, particularly the old Khan laboratory. The labs have always been a black hole for American officials, a no-go zone. There, and at a competing weapons development facility run by Pakistan's Atomic Energy Commission, work is under way to speed the production of a new generation of bombs, smaller weapons utilizing plutonium. With each succeeding generation of nuclear weapons development, there is more and more knowledge circulating about how to build a bigger bomb with less material. And while you can lock down weapons with codes and fences and guards, locking down expertise is a lot harder.

Working in some secrecy, the Bush administration attempted, laudably, to help Musharraf with the nuts and bolts of nuclear security. The effort was contained in a highly classified program that started in the months immediately after 9/11, when people began to realize that the terrorist attack, horrific as it was, could have been a lot worse.

The problem is that, years later, no one is sure how well the program is working. Like everything else that has happened in six years of an uneasy alliance between America and Pakistan, the effort to secure the weapons has been undercut by mutual distrust. Publicly, American officials almost never talked about Pakistan's nuclear weapons program. Privately, many inside the White House were obsessed with it. Once, early in 2007, Bush mused openly about his fears regarding the Pakistani intelligence service. He said he believed the ISI never really cut its “old school ties” with the Taliban.⁷ Months later, as Pakistan appeared on the verge of political meltdown and Musharraf was slipping toward irrelevancy, one of Bush's nuclear experts acknowledged to me that “if the place falls apart, all the assurances we've been given about the safety of the weapons are worth squat.” (Others, notably Bush's national security adviser, Stephen Hadley, disagreed.) The fact is, the Bush administration didn't know what would happen to the nuclear weapons if Pakistan dissolved into chaos, and neither will Obama's.

YET EVEN IF the United States had its priorities in order, securing the world's bomb fuel and loose weapons would not be easy, as I learned after visiting Kidwai. In the new nuclear age, our allies—the Pakistanis chief among them—distrust our motives just as our enemies did during the Cold War. And so, when he heads to work inside Chaklala Garrison these days, Kidwai has several preoccupations. One is stopping the outsiders who might attack Pakistan's nuclear facilities, a group that includes everyone from the Indians to al Qaeda to brutal new tribal leaders such as Baitullah Mehsud, the thuggish militia leader whom the CIA believed ordered the Bhutto killing. The second is stopping insiders who might be planted in the program to seize part of Pakistan's arsenal, or maybe just its bomb-making secrets. But his third problem is us.

Kidwai is nothing if not a realist. He knows that Washington's offers of technological aid are both a blessing and a threat. He has no doubt that the United States has developed extensive contingency plans to seize or neutralize Pakistan's nuclear weapons if it fears that terrorists might get to them. The truth, American officials tell me, is that those “plans” are still more like hopes. “It would help,” one official said, “if we knew where the damn stuff was.”

The result of these deep mutual suspicions is that the United States knows little about what's happening inside Pakistan's nuclear complex. “It scares the hell out of me,” said a senior Bush official who spent a lifetime studying the Soviets. “Every morning I could see what was happening inside the Soviet nuclear system. I've never had a morning when I could see inside Pakistan's.”

That explains why Kidwai and his Strategic Plans Division have grown ever more central to Washington's strategy, however incomplete, about what to do if the only nuclear-armed nation in the Islamic world begins to slip over the precipice. For all the public talk about democracy and development, about the need to foster moderation in Pakistani society, in the end it is the security of that arsenal that captivates Washington's attention.

KHALID KIDWAI is only a few years younger than Pakistan itself, and he has spent much of his life trying to create pockets of order in a nation to which order does not come naturally.

In his youth, Kidwai was protected from these Hobbesian elements of Pakistani life. He grew up in a world of Urdu literature and poetry. The enemy was India, not forces within Pakistan itself.

His father, Jalil Ahmed Kidwai, who lived to be ninety-two, was one of the country's best known authors and critics, whose studies of Indian, Pakistani, and Western literature are still the stuff of Ph.D. dissertations. His mother was a school principal in Karachi, the now-violent port city on the Arabian Sea where the family moved during the turbulence of the partition from India in 1947. The central question on the minds of most Muslim exiles at this time was whether Pakistan could withstand India's onslaughts, and it did not take long for the young Khalid to settle on his dream: to fly in the Pakistani Air Force, the most romantic branch of the armed forces of a new nation that needed to be able to strike deep into India if it was to survive.

At age twelve, he passed the exam for the Air Force-sponsored school in Sargodha, the site of the country's largest air base. “I wanted to make it as a pilot,” he told me, “and it was the fastest way.” Three years later, as a tenth-grader, he got a firsthand view of what war with India would look like. On September 6, 1965, he and his classmates were hurriedly sent home; war had broken out, and the students were in the bull's-eye. The next day Indian warplanes attacked the Sargodha base, but to little effect. When the students fled, most of the Pakistani aircraft based there were quietly moved elsewhere.

Sargodha remains a defining part of Kidwai's life, largely because much of Pakistan's nuclear arsenal is stored there or nearby. But Sargodha also is the place where his early dreams were dashed. When he graduated, Kidwai received the disheartening news that he would never become a pilot; a mild disorder with his eyes disqualified him. “My next obvious choice was the army,” he said, and like many in his generation of military men in Pakistan, he stayed on, enjoying the professional pride and the security blanket it provides.

In 1971, as a young second lieutenant, Kidwai was suddenly in the thick of two vicious wars. The first was the civil war with what was then called East Pakistan, today known as Bangladesh, during which he was promoted to captain in an artillery regiment. On November 23 of the same year, war broke out again with India, and Kidwai was captured and held as a prisoner of war for two years in Allahabad, India. It was an experience he is reluctant to talk about.

After his release, the army made sure Kidwai was exposed to the two countries with which Pakistan had its most crucial relationships: the United States and Saudi Arabia. In 1979, he was posted to Fort Sill, Oklahoma. For a newly married young officer, still in recovery from his years in prison camp, it was a jarring exposure to a world that could not have been more different from his own.

Fort Sill is the last of the old Indian forts, created by a hero of the Civil War, Maj. Gen. Philip H. Sheridan, to finish off the Comanche, the Apache and others threatening the Western settlement of the country. Geronimo lived there, along with more than three hundred other Apache prisoners of war, and died there a century ago. But by the time Kidwai arrived, Fort Sill had long since reinvented itself as the U.S. Army's artillery school and a place where young officers of allied nations came for training. It offered a program—later suspended during America's effort to isolate Pakistan following its nuclear tests— that allowed Americans to get to know a rising generation of Pakistani officers. And it gave Kidwai an appreciation of the United States that is still somewhat rare among the Pakistani elite. He talks longingly about life in Oklahoma's wide-open wheatfields, of traveling through Colorado and Texas, and even of witnessing the awesome power of a tornado that wiped out a whole neighborhood of Lawton, Oklahoma, where he and his new wife were living. But what really struck him was the warmth of the local families who sponsored them, despite the paucity of Muslims in Comanche County.

It was also at Fort Sill that he caught his first whiff of the role of nuclear weapons in the modern military. The artillery school at the fort prepared American officers for Cold War operations in Europe and regularly drilled them on the procedures under which tactical nuclear munitions might be used to fight off a Soviet force invading Western Europe. For Kidwai, this seemed something of a remote subject; India had just tested its first bomb, and Pakistan was still far from obtaining a nuclear deterrent. But the subject intrigued him—chiefly because whenever the nuclear training became intense, the brass at Fort Sill found something else for the foreign officers to do.

“We'd be sent off for trips to Washington or someplace,” Kidwai recalled with a laugh, “so that we were out of earshot.”

As he worked his way up the Pakistani Army's rigid promotion system, he found himself posted next to Saudi Arabia. There he gained a view of a very different kind of ally for the Pakistanis—a Sunni nation so awash in oil money that it could pay Pakistan handsomely to help defend the country.

For roughly two years Kidwai lived along the Jordanian border, while his wife learned Arabic by socializing with ladies from the region. By May 28, 1998—the day on which Pakistan's power and Kid-wai's life would change dramatically—he was back in Pakistan, based just south of Lahore, an eager brigadier general just days away from his promotion to major general.

Even today, you can see Kidwai's pride in the scope of Pakistan's accomplishment that day as the Chagai Hills shook from Pakistan's first underground nuclear test. His country had done more than answer India's challenge; it had built the ultimate deterrent. Along the way, Pakistan had overcome a series of halfhearted efforts, led by the United States, to cut off its nuclear supplies and dissuade it from building its own bomb.

Year after year Pakistan lied to Washington when confronted with all-but-definitive evidence that it was constructing a weapon. When Washington could no longer overlook the obvious, Pakistan simply endured the resulting economic sanctions, even though the country's economy was flat on its back. It all seemed worth it, Pakistani officials have told me, after that first test detonated without a hitch. Then, just for good measure, the military detonated five more tests over the next few days, with Khan present at the last test to accept the kudos, even though others had been the ones to design the weapons.

Kidwai told me there was a special satisfaction in the fact that Pakistan had conducted six successful tests in those mountains. “That was one-upsmanship,” Kidwai said, smiling proudly as we looked at a photograph of one of the tests hanging on his office wall. “India had conducted only five.” Underneath the photographs, Kidwai keeps a small fragment of the Chagai mountain under glass, displayed like a moon rock at the Smithsonian. The explosion had turned it bright white.

Kidwai professes not to have known in advance about the tests. “I had no clue,” he said, “because I was not a part of the nuclear establishment at that point.” In fact, the tests were so sensitive that very few in the upper echelons of the military were clued in that tests were on the way. But it was no secret that the army, the keeper of the weapons and the institution with the executive authority over their use, was determined to demonstrate that Pakistan had the bomb and would never give it up, no matter what the diplomatic and economic costs.

India's test had given the army, and Khan, the political cover they needed to demonstrate their technical accomplishment. Each got something out of the show. Pakistani authorities could claim that they were merely responding to a provocation from the Indians, who had gone first, in hopes they would escape sanction. But the Pakistanis also relished the opportunity. “You have to understand there were huge sensitivities,” Kidwai said. “There is a history of the world trying to roll it back, to stop us, to get us to give it up.” Khan cemented his role as a national hero, and soon his laboratory was swarming with North Koreans, clearly interested customers.

The Clinton administration was helpless, but made one last, halfhearted effort to force Pakistan to stuff the genie back into the bottle. Angry because he had put his credibility on the line in pressing Prime Minister Nawaz Sharif not to conduct the tests, Clinton imposed heavy sanctions on both India and Pakistan, as required by a law called the Glenn Amendment, designed to make any new nuclear power pay a price for conducting a nuclear test. Clinton declared that Pakistan had missed a chance to “improve its political standing in the world,” and said both India and Pakistan had to “take decisive steps to reverse this dangerous arms race.”⁸ Six billion dollars in aid was canceled, and lending from the World Bank and the International Monetary Fund was delayed.

While they protested, both Indian and Pakistani officials knew that nothing lasted forever—especially sanctions. In fact, just after September 11, 2001, three years after the tests, the Bush administration suddenly concluded it needed Pakistan as an ally in the war against the Taliban and al Qaeda. The sanctions were lifted overnight. The message to the rest of world—Iran included—was clear: Hang tough, proceed with your nuclear ambitions, and sooner or later the West will need you badly enough to accept you as another nuclear power.

No sooner had the radioactive and diplomatic dust settled from the test site than Kidwai was called in by his boss, the chief of the army staff, Gen. Jehangir Karamat. A former armored corps commander and staunch nationalist, Karamat was a veteran of both Indo-Pakistani wars, in 1965 and 1971. He had commanded the Pakistani troops in Saudi Arabia, and tangled with Prime Minister Nawaz Sharif over his insistence that the military be given a clear role in running the country. He told Kidwai that he had to put everything else aside for an urgent project: to come up with a system to protect Pakistan's new atomic weapons from all enemies—the Indians, the West, and especially the angry Americans.

Together, Karamat and Kidwai knew speed was of the essence. If the West thought that Pakistan had a few weapons in its inventory, and no system to keep them safe, Pakistan's leaders feared they would come under even more pressure to “roll back” the program and give up the handful they had manufactured. The only way to resist that pressure, they knew, was to create a large arsenal fast and to hide it away in tunnels and caves where neither the Indians nor the Americans could seize or destroy the warheads.

Moreover, Islamabad needed to convince the world it could become a responsible nuclear power that was just as capable of securing its weapons as the Russians, the Chinese, or the Israelis. That was where Kidwai and his team came in.

Kidwai is the first to concede that he came to the job with zero expertise. Apart from his limited exposure to nuclear artillery issues at Fort Sill, he had no real experience in the subject. He had no scientific background in the design or manufacture of nuclear weapons, or in the mechanisms to lock them down. He had never studied deterrence theory. None of this seemed to matter. He was considered a quick study and, most important, a loyalist. By October 1998, he had developed the rudiments of a plan, and he presented it to Musharraf, who was just months away from taking over the country in a military “counter-coup” against the prime minister he detested, Nawaz Sharif.

But as Musharraf later told me, he and Kidwai quickly came to realize that the greatest danger to their plans for nuclear control lay not in New Delhi or in Washington. It was a short walk down the road, at the Khan laboratory. They just didn't know what to do about it.

ABDUL QADEER KHAN had not been present at the creation of the Pakistani nuclear program, but he insinuated himself into it with astounding speed.

Pakistan's quest had begun more than two decades prior to India's first nuclear test in 1974. Long before coming to power as Pakistan's military leader, Benazir Bhutto's father, Zulfikar Ali Bhutto, had dreamed of a Pakistani nuclear weapon, arguing that “all wars of our age have become total wars.” He toured the Muslim world to raise money. He stopped in Iran, Saudi Arabia, and eventually Libya, to see Col. Muammar el-Qaddafi, who harbored similar dreams. Soon the money began to flow, a reported $100 million from Qaddafi. The Saudis are also believed to have donated heavily, but to this day top American officials say they have never seen definitive evidence about what they got in return. It didn't take a nuclear scientist to figure out their potential motives: Worried about the Iranians, the Saudis are presumably eager to have access, if needed, to what some call “the Sunni bomb.”

The Indians, however, had a head start. In May 1974, with Indira Gandhi as a witness, they had conducted their first nuclear test. The Indians called it a “peaceful explosion,” but because the test was conducted within a hundred miles of the Pakistani border, the message was short of subtle. “I never really understood what kind of ‘peaceful explosion’—a nuclear explosion, how can it be peaceful?” President Pervez Musharraf said to me during a conversation in 2005. To Musharraf, a young officer rising through the army ranks at the time, the Indian test left Pakistan so vulnerable that “our strategy of minimal deterrence was undermined, was compromised.” It did not take long, he said, before “we decided that we had to go nuclear.”⁹

Years later, Benazir Bhutto recalled during a conversation in London that her father was overcome with shock and shame when he first heard about the Indian test. He said privately to her, and then publicly, that “we will eat grass, but we will build the bomb.”¹⁰ For a desperately poor country, making the declaration and accomplishing the goal were two very different things. Bhutto was told that designing a bomb was not very hard. Coming up with the fissile material to make one was an entirely different matter.

The message raced through the military. It was then, in a case of spectacular timing, that Khan wrote a letter offering up his services from the Netherlands, where he was working at a European consortium that produced centrifuges for uranium enrichment. His offer quickly reached the upper echelons of the Pakistani military. “For Pakistan after 1974 there was nothing more important than the bomb,” Talat Masood, a retired senior general in the army told me one evening as he was serving coffee in his living room.¹¹ Khan was soon back home, his files stuffed with stolen production documents. Soon he had his own lab, which he modestly named for himself.

“The two laboratories, the Khan Research Lab and the Pakistan Atomic Energy Commission, were given an unprecedented level of freedom and resources considering the poverty level of the country,” Masood told me. “All the security around them was not to make sure that they didn't abuse the authority; it was to protect what they were making.”

While the Pakistanis saw the bomb program as part of their epic competition with India, to the rest of the world it marked the fruition of John F. Kennedy's prediction that many more states would soon become nuclear powers. Suddenly, the great powers— the United States, Russia, Britain, France, and China—no longer held a monopoly on nuclear weapons. India's “peaceful explosion” and Pakistan's reaction were bound to change the world. They did, but not in the ways many anticipated.

Year after year, as American presidents and Pakistani prime ministers came and went, intelligence agents from the United States and Europe watched as Pakistan slowly gathered all the elements it needed for a bomb. High-performance metals from Europe and nuclear triggers from the United States, at least until the operation to send them through third countries to mask their true destination was shut down, were being shipped to Pakistan. Though American intelligence agencies missed it at the time, there was also the nearly complete bomb design from China, the one that ended up in Libyan hands decades later.

Yet over the years Democratic and Republican presidents had averted their eyes from the overwhelming evidence being gathered by the CIA. The reason was simple: Washington needed Pakistan in the war against the Soviets in Afghanistan and later to support the forceful removal of the Taliban. To declare publicly that Pakistan was building the bomb would have required Washington to cut off aid to the country, something the White House believed it could not afford to do. So, in time-honored tradition, it fudged the evidence, neither denying nor confirming in public the proof it had in hand.

For its part, the Pakistani military—and Khan in particular-operated with equal secrecy. Before her death, Benazir Bhutto insisted that when she became prime minister in December 1988, her own commanders refused to discuss the project with her. The army chief, she said, had told her, “There's no need for you to know.” He was clearly concerned that she would give in to pressure from the Americans, her patrons.

But to the CIA, Bhutto looked too weak to confront the Pakistani Army on a major issue of national security. To make the point that they knew what Pakistan was up to, Bhutto was escorted in June 1989 to CIA headquarters in Langley, where she was presented with a full American mock-up of the Pakistani bomb and warned that the United States was following every step. She realized from that experience, she argued later, that she was being kept in the dark about the state of the weapons program. But she also contended that the Americans were helping her come up with subterfuges that would enable all sides to talk around the truth. “We were told that if we wanted to keep the aid” from the United States, she said in 2005, Pakistan should simply avoid assembling all the parts into a completed weapon. “When is a chicken a chicken?” she asked in the same interview, describing this diplomatic sleight of hand with a smile.

While Washington was quietly teaching Bhutto the art of nuclear ambiguity, it was missing the big turn in Khan's program. He was double- and triple-ordering critical parts for Pakistan's program, then repacking and selling some of the parts and the technology. “He knew there was no accountability in the system,” Talat Masood recalled. “And why should he? He was at the center of attention, he was highly egotistical, he was highly ambitious,” later making it clear he thought he should be running the country.

He was also somewhat delusional about Pakistan's own capabilities. General Masood remembered confronting Khan about the security of missiles he was buying for the Khan laboratory. Khan insisted he was building the Ghauri, one of Pakistan's first nuclear-capable missiles, from his own parts and designs. “It was ridiculous, because everyone knew we were buying these from the North Koreans, and we had to be giving them something in return.” Khan, he noted, soon wanted to produce conventional weapons as well—antitank missiles, shoulder-fired antiaircraft missiles—which investigators believe he wanted to trade for nuclear technology. But he may also have simply wanted to sell them.¹²

During her exile in London, Bhutto insisted that she never knew that Khan was seeking to profit by selling Pakistan's technology. But she said she had “noticed a big change between Khan in my first term and Khan in my second term.” In her first term, she said, he had been “very nationalistic, very proud” as he marched in and out of meetings declaring that he was the man who would match India, bomb for bomb, missile for missile. Clearly he had more influence over the program than she did: Bhutto claimed that during her two terms as prime minister, she was not in command of the nuclear weapons program. The first time she was removed from office, in 1990, she claimed she was a victim of a “nuclear coup,” triggered in part by her insistence that she had the right to convene the National Command Authority, the government organization set up to decide Pakistan's nuclear doctrine and control all major decisions about its nuclear weapons—including when to use them.

When she returned to power in 1993, she once again found herself unable to extend her reach over the nuclear program, or so she said. Meeting Khan again, she recalled, she saw a changed man: the nationalism had given way to a more religious tone.

“I had noticed that he had become more Islamist,” she said. Khan kept talking about his mission to create an “Islamic bomb” that would match the “Jewish bomb.” In retrospect, he may have simply been creating a rationale for spreading the technology around the world, one more publicly acceptable than pocketing tens of millions of dollars in profits.

Bhutto laughed when she heard the Pakistani government's claim that top officials did not know that Khan had shifted gears and begun selling his technology. “I do not buy this theory that Khan did everything on his own,” she said. “But maybe it suits Mr. Musharraf to say Khan did everything on his own.”

It also suited her to blame Musharraf, the rival she was trying to take down. When it came to aiding and abetting Khan, however, her own hands were not entirely clean. On a visit to North Korea in the 1990s, according to accounts provided by Khan's associates and by American intelligence officials, Bhutto reportedly brought back to Islamabad some North Korean missile designs that would enable Pakistan to ensure that its nuclear weapons could reach deep into India. North Korea later got uranium enrichment equipment, according to statements Khan made after his arrest in 2005. It is still unclear what was quid and what was quo. Until her death, Bhutto remained maddeningly evasive about her role in the exchange.

IF THERE IS a single story out of the Khan laboratories that sends a shiver down the spines of American officials, it is the tale of Sultan Bashiruddin Mahmood.¹³

Until 1999, when he was quietly removed from the Khan laboratories, Mahmood was one of the scientists who worked in utmost secrecy on the gas centrifuge program that Khan stole from the Netherlands and brought back to Pakistan. Mahmood then moved on to the country's next huge project: designing the reactor at Kushab that was to produce the fuel Pakistan needed to move to the next level—a plutonium bomb.

Mahmood grew up in India, and, after the 1947 partition, moved to a village outside of Lahore, where his family lived in poverty. He was among the lucky few who won a scholarship to college and, by the 1960s, found himself in Manchester, England, working on a degree in nuclear engineering.¹⁴

He returned to Pakistan after six years of study and quickly became a student of the intersection between fringe science and the Koran. He wrote The Mechanics of the Doomsday and Life after Death, a book arguing that in parts of the world where morals degrade, disaster strikes. Though he was acknowledged as a highly capable engineer and scientist, over time his colleagues began to wonder if Mahmood was playing with less than a full deck. He talked often of his fascination with sunspots, and his next treatise was an extensive essay in Urdu about the role sunspots played in triggering the French and Russian Revolutions, World War II, and uprisings against colonial masters around the world.¹⁵

“This guy was our ultimate nightmare,” an American intelligence official told me in late 2001, when we first wrote about Mahmood. “He had access to the entire Pakistani program. He knew what he was doing. And he was completely out of his mind.”

While Khan appeared to be in the nuclear-proliferation business chiefly for the money, Mahmood made it clear to friends that his interest was religious: Pakistan's bomb, he told associates, was “the property of a whole Ummah,” referring to the worldwide Muslim community.¹⁶ He made no secret of his goal: He wanted to share nuclear technology with anyone who might speed “the end of days,” which he said would pave the way for Islam to rise as the dominant religious force in the world.

Eventually his religious intensity, combined with his sympathy for Islamic extremism, scared Mahmood's colleagues. In 1999, just as Kidwai was beginning to examine the staff of the nuclear enterprise, Mahmood was forced to take an early retirement. At a loss for what to do, he set up a nonprofit charity, Ummah Tameer-e-Nau, ostensibly designed to allow him to send relief to fellow Muslims in Afghanistan. It turned out that the ones most interested in Mahmood's efforts were Osama bin Laden and his deputy, Ayman al-Zawahiri, the former surgeon from Egypt who had become the chief operational officer of al Qaeda.

In August 2001, as the September 11 plotters were making their last preparations in the United States, Mahmood and one of his colleagues at the charity, Chaudiri Abdul Majeed, met with bin Laden and Zawahiri over the course of several days. Years later, what exactly transpired in those meetings remains a mystery. There is no doubt that Mahmood and Majeed talked to the two al Qaeda leaders about nuclear weapons, and there is no doubt that al Qaeda desperately wanted the Bomb. George Tenet, the CIA chief, recalled later that reports of the meeting were “frustratingly vague,” but included an account of how a senior al Qaeda leader displayed a canister that may have contained some nuclear material, and there was talk of how to design a simple firing mechanism.¹⁷ What no one knew is whether this session was just about bin Laden's wish list, or whether al Qaeda had a plausible plan to obtain a nuclear weapon.

Either way, the flood of intelligence that poured into Washington in the days after 9/11 set off a panic that the attack, horrific as it was, might be just a start. Mowatt-Larssen, the longtime CIA nuclear expert, now at the Energy Department, was given perhaps the most daunting job at the agency in the aftermath of 9/11: to make sure that al Qaeda did not have a worse weapon at its disposal. “The worst nightmare we had at that time was that A. Q. Khan and Osama bin Laden were somehow working together,” recalled Mowatt-Larssen. “There were all these connections, and connections to the Libyans,” who reported to the British and others that they had been approached by bin Laden's group with offers to sell a weapon.

In retrospect, this was the moment when Bush began to shape his foreign policy demands that foreign leaders make countering terrorism their highest priority, a loyalty test that was understandable at the time but eventually came to skew, and often undermine, America's interactions with the world. When Colin Powell landed in Pakistan in mid-October 2001, Ground Zero was still smoldering, the American invasion of Afghanistan was looming, and Powell needed to bring answers back to Washington—fast. Publicly, Powell's trip was designed to stress a new era of cooperation with Musharraf and “to demonstrate our enduring commitment” to “a great Muslim nation.”¹⁸ But its true purpose was to offer more help to Pakistan in controlling its own arsenal and to confront the Pakistani leader with evidence to force him to detain Mahmood and Majeed, and allow the CIA to question them.

Facing no alternative, Musharraf had the two arrested six days later. But after a cursory interrogation—the suspects were permitted to go home at night—they were released. Then, in November 2001, Tenet and Mowatt-Larssen briefed Bush and Cheney about the campfire meeting with bin Laden. For the first time, Mowatt-Larssen recalled, “we were looking at the idea of terrorists looking at full strategic weapons, the kind of weapons only states had.” The evidence suggested, but did not prove, a connection between al Qaeda and the Khan network, perhaps with the involvement of the Libyans. This collaboration among nuclear aspirants later proved to be less than met the eye. But at the time it could not be dismissed. Bush and Cheney found the prospect so alarming that they told both Tenet and Mowatt-Larssen to head for Islamabad immediately to confront Musharraf about the meeting with bin Laden. Within hours, Tenet and Mowatt-Larssen were on an aging Boeing 707, used decades before as Air Force One, and the next day they were ushered into Musharraf's office.

“The reaction was interesting,” Mowatt-Larssen recalled later. “Musharaf's first words were ‘Men in caves can't do this.’” He told Tenet and Mowatt-Larssen they were looking for nukes in all the wrong places; why didn't they focus on Russia and the former Soviet Union? After an hour of intense conversation, Musharraf came around and ordered the two rearrested. But while the Americans could join the interrogation, he said, Pakistan could not acknowledge their presence. Musharraf could not be perceived as caving in to U.S. demands.

At first Mahmood denied ever meeting the al Qaeda leaders. Then he failed a series of polygraph tests, administered by Americans who had slipped quietly into the country. Eventually he and Majeed were confronted with evidence showing that they had promised to help bin Laden build a weapon. Gradually their memories improved. Mahmood described the meetings, bin Laden's persistent questions about what it would take to obtain a bomb, and his explanation that building the weapon would be a lot easier than obtaining the right kind of fuel. Bin Laden, he reported, asked, “What if we already have the material?”

Bin Laden's single previous effort to obtain that material turned out to have been a failure. The al Qaeda leader apparently fell for a scam out of Uzbekistan that provided him with some mildly radioactive material that might have been of use in a dirty bomb, but not in a nuclear weapon. (During their visit, Mahmood and Majeed were reportedly shown a canister by a senior al Qaeda leader that they were told contained some nuclear material, but it seems likely it was low-level stuff, perhaps the expensive fruits of the Uzbekistan rip-off.)¹⁹

Under interrogation, Mahmood and Majeed said the conversation left them with the impression that al Qaeda already had a nuclear program under way, albeit a haphazard one, in which desire outstripped supplies of fissile material and expertise.

The interrogations of Mahmood and Majeed were deeply unsatisfactory. For all his desire to give jihadists a power akin to Pakistan's, Mahmood was a specialist in the fuel-production side of the business, enriching uranium and making plutonium. He had no access to nuclear material and had never designed a bomb. Mah-mood's son, talking to reporters later, said that his father had explained to bin Laden that there was nothing easy about building a bomb and that al Qaeda would first have to obtain nuclear fuel.

Still, for the CIA, fresh from the failures leading up to 9/11, it was scary stuff. Tenet's description of the evidence to Bush and Cheney prompted the vice president to mumble his “1 percent” rule: that even if the chances of a nuclear 9/11 were only 1 percent, the possibility had to be eliminated as if the evidence were overwhelming. Tenet later argued that Cheney's statement was “misinterpreted” as a policy to pursue all threats, no matter how remote, as if they were certainties. But Cheney was right about one thing: Nuclear terrorism is different. It is, as Tenet wrote later, one of the few attacks terrorists could mount that would “change history”²⁰

In the end, Mahmood and Majeed were never prosecuted because the Pakistanis did not want to risk a trial in which the country's own nuclear secrets could come out, much less broadcast the carelessness that had allowed one of their former scientists to meet unnoticed with bin Laden. Today, Mahmood is back home, under tight surveillance that seems intended more to keep him a safe distance from reporters than to keep him away from extremists.

Kidwai insists that Pakistan's investigations have concluded only that Mahmood and his friends engaged in “some sort of discussions” about how to make nuclear weapons and that bin Laden “perhaps showed some interest” in how to build them. Mahmood had probably drawn basic sketches of a nuclear weapon, he said, but they were crude drawings, more like something you would scribble on a napkin, than a detailed plan. Kidwai dismisses the whole matter as the misadventures of a wayward scientist.

In Kidwai's telling of events—a version of the story that is a lot more benign than the version heard in Washington—”nothing went anywhere.” He clearly wants this episode to be dismissed as revealing bin Laden's wish list, a list that he argues was never fulfilled. “It's over,” he insists.

Still, one of his deputies acknowledges that the meetings with al Qaeda “will haunt us for many years” because they will fuel suspicions around the world that Pakistan is unable to control its most precious asset: the nuclear knowledge that exists inside the heads of its scientists and engineers.

THE TERRIFYING PART of Mahmood's story is not just what transpired around the campfire, but rather that the meetings happened at all. They took place three years after the Strategic Plans Division came into existence, and demonstrated the huge vulnerabilities in the Pakistani nuclear infrastructure.

By Kidwai's count, there are roughly 70,000 people who work in the country's nuclear complex. Roughly half of those are involved in the technology of the program, including 7,000 to 8,000 scientists. Of those, Kidwai estimates that about 2,000 are “hard-core nuclear scientists and engineers” with critical knowledge of how to build a weapon. They are the focus of the Strategic Plans Division's personal reliability program, an effort that Kidwai characterizes as a screening program for the medically unfit and psychologically unreliable. Intelligence experts in Washington describe the program in far starker terms, saying they hope it weeds out religious zealots, Taliban sympathizers, al Qaeda spies, and the economically desperate, ensuring that they are not in the ranks of Pakistan's nuclear hierarchy.

The differences in description are worrying, though the simple explanation may be that Kidwai does not want to admit that there are elements of Taliban and al Qaeda looking to get inside the nuclear program. When I asked one of Kidwai's Americanized aides to describe the program, he said, “We're looking for the fundos [fundamentalists], and that is not always easy to figure out.”

At its heart, the personal reliability program is about what we would call domestic spying: Kidwai has created his own intelligence service, so that he does not have to rely entirely on the ISI, an agency that is believed to be infested with Taliban sympathizers. The intelligence arm of Strategic Plans monitors everything from bank transactions to religious habits to the political persuasions of the country's nuclear engineers. When Musharraf was still at the height of his power, an engineer was dismissed, allegedly because he was overheard speaking ill of the president and his policies.

But the truth is that weeding out “fundos” in Pakistan is hard work. Unlike Mahmood, they do not necessarily publish bizarre theories. Pakistan is, after all, an Islamic state, and it would be a dangerous leap to assume that any fervent Muslim would sell out the country's nuclear arsenal. (In fact, one retired Pakistani general told me that A. Q. Khan would have likely passed the personal reliability test with flying colors, unless his overseas bank accounts were monitored.) Like the borders between Pakistan and Afghanistan, the line between religious fundamentalism and militant fundamentalism is hard to draw.

A visit to Pakistan's leading university campuses reveals that something has changed. A decade ago, most young women walked the campuses with their heads uncovered. Not today. “I look out at my classes today, and among the women I can't see many of the faces,” Pervez Hoodhboy, one of the lonely critics of the Pakistani nuclear program, told me when I visited him at his office at Qaid-e-Azam University in Islamabad in the spring of 2008. Intense and prone to question authority, Hoodhboy is chairman of the physics department.

“That was rarely the case before,” he said, showing me a photo of some of the women in his class twenty years ago who were dressed in blue jeans.

A craggy-faced scientist who spends most of his time sparring with Pakistan's religious and nuclear establishments, he has the air of a rebel resigned to the fact that he will always be regarded as an outcast in his own country. Almost alone, he has argued that Pakistan's nuclear weapons program has made the country less safe. His critics, including Kidwai, tend to cast him as the physicist who cried “fundo” once too often. But Hoodhboy has a good record of documenting undercurrents of Pakistani society that the government, in its effort to portray a moderate face to the world, would rather suppress.

His argument is simple: As the nuclear program attracts new, young talent, it will be next to impossible to weed out militant fundamentalists. What's more, the country is turning a blind eye as fundamentalism is taught in the schools—not just in the religious madrassas, but in the public and private schools that educate the rest of the population. To prove his point, Hoodhboy showed me pictures he has collected from a schoolbook used by both regular schools and the radical madrassas to teach the Urdu alphabet. To American eyes, it is more than a little shocking. The problem isn't the words; it's the pictures.

Take the word collision, which is pronounced tay in Urdu. The illustration? A picture of two airplanes flying into a burning World Trade Center. “Of all the pictures of a collision that you could find,” Hoodhboy told me, “it's curious that they use that one in the textbook.”

Naturally, members of Musharraf's government have assured visiting Americans that the textbooks are being cleaned up and that sections advocating jihad are being deleted. Hoodhboy makes a persuasive case that the changes have largely been for show. What's more, a $100-million government project for “madrassa reform” was killed after it met with enormous objections from the religious establishment.

And the problem is not limited to basic readers for first-graders. Leafing through today's high school textbooks is a little like reading Japanese textbooks about World War II, the ones in which the Rape of Nanking is still described as an event that may or may not have happened. In Pakistan's case, the textbooks say little about Afghanistan or the brutal rule of the Taliban; they skip right ahead to Iraq, arguing that Americans occupied the country to seize territory and oil. Some suggest Washington has designs on the rest of the region. However twisted, it is a storyline that George Bush made far easier to sell.

THE PANIC IN WASHINGTON about what al Qaeda sought from Pakistani scientists led to a sea-change in the American approach to the country: Instead of sanctioning Pakistan for possessing nuclear weapons, America began paying Pakistan to lock them down. It was a program shaped by the very public efforts to lock up nuclear weapons and material in the Soviet Union. Washington boasts about that project, but to this day officials go silent when asked about the parallel program in Pakistan, for fear of destabilizing the government.

Bush never engaged in the issue prior to 9/11. But after the attacks, the administration began acting like parents who discover, belatedly, that one of their kids has been having sex: It dropped the lectures about the virtues of abstinence and started talking about safety.

When Colin Powell traveled to Pakistan in October 2001, one of the real purposes of what he called one of his “general-to-general” conversations with Musharraf was to impress upon him that whatever Pakistan was doing to secure its arsenal, it wasn't enough.

Powell knew what he was talking about: As a young Army officer in Germany during the Cold War, he was a “nuclear weapons employment officer,” which meant, he told me, that “I knew how to sit down with various maps and charts, and describe where to put a nuclear weapon and what it would do to the enemy.” It was an experience that also gave him a hands-on sense of how easy it might be, in a moment of chaos, to lose track of where a weapon is located.²¹

He told Musharraf it was urgent that he accept American help in locking down his program. Powell knew that back in Washington, where paranoia about a second attack was through the roof, one nightmare scenario after another was being played out. In some, bin Laden exploited connections or bribed someone to get his hands on a real weapon. In others, the ISI—long sympathetic to the Taliban—helped its friends get their hands on the one weapon that could stop the coming American-led invasion.

“There were a lot of people who feared that once we headed into Afghanistan, the Taliban would be looking for these weapons,” one senior official involved in the effort told me.

Musharraf later had another interpretation of events: The Bush White House thought he was at high risk of being killed or deposed, and they feared a free-for-all for the weapons. In his memoir, In the Line of Fire, Musharraf recalled being “put under immense pressure by the United States regarding our nuclear and missile arsenal.”

“They were not very sure of my job security, and they dreaded the possibility that an extremist successor government might get its hands on our strategic nuclear arsenal. Second, they doubted our ability to safeguard our assets.”²²

But when I talked to him in 2005, Musharraf argued that he could not afford to appear to be giving in to American pressure. Any hint or rumor that he was allowing American hands to be put on Pakistan's one great source of power would be a death knell for any Pakistani leader.

“This is an extremely sensitive matter in Pakistan,” Musharraf said. “We don't allow any foreign intrusion in our facilities. But, at the same time, we guarantee that the custodial arrangements that we brought about and implemented are already the best in the world.”²³

Kidwai was even blunter. “Powell made the offer, and we had no problem in agreeing to examine that,” he recalled. “But we were highly sensitive to the local sentiment that the slightest interaction with the Americans would be seen as a sellout.”

The result was that Bush and Powell boasted in public about their efforts to secure nuclear weapons and materials in Russia, and ordered enormous secrecy around the parallel effort in Pakistan. They never spoke about the program in public. The budgets, though modest, were highly classified.

When I began asking around the White House about the program, the president's national security adviser, Stephen Hadley, asked me to come to his office to hear his arguments for why the Times should not publish stories about the program. Premature publicity, he argued, could undermine the effort to secure the weapons at a time when many of them had probably not been properly locked down. Moreover, he said, the revelation could galvanize opposition to President Musharraf and fuel the argument that he was an American lapdog.

Hadley's first argument, that a story would aid terror groups, convinced the Times's editors to delay publication until the protections were better established. (At the same time we told the administration that we do not withhold stories simply because they would prove embarrassing to American or foreign officials.) Ultimately the story was held for three years, until the chaos in Pakistan in the fall of 2007 urgently raised the question of whether the nuclear stockpile might be vulnerable during a potentially violent overthrow of Musharraf. (When I told the White House that the events in Pakistan were leading the paper to revisit our earlier agreement to delay publication, Hadley withdrew his request, and within days the Times published many of the details in a front-page story.)^1*

Yet when the program began, Powell knew there were risks that if the secret of the American aid to Pakistan was kept too close, India might be tempted to act on its own to destroy or seize Pakistan's nuclear stockpile, thereby sparking another conflagration. So in the fall of 2002, at a time of renewed tension between India and Pakistan, Robert D. Blackwill, the U.S. ambassador to India, got an urgent message one weekend from Powell. He was to go to see the prime minister of India, Atal Bihari Vajpayee, immediately. A phone call would not do, nor could it wait until Monday. He was to read Vajpayee a brief statement explaining that the United States was providing assistance to the government of Pakistan to help secure the country's nuclear weapons. Blackwill was given explicit instructions to answer no questions, to leave no paper with Vajpayee, and to insist on total secrecy.

“It was pretty comical,” Blackwill recalled later about the instructions. Vajpayee's foreign policy advisers would not allow the aging prime minister to be completely alone with the American ambassador, especially one with Blackwill's fearsome reputation and skills. Blackwill lived within the letter of his instructions and left no paper—but, at the insistence of the Indians, he allowed them to take extensive notes on what was said. Years later, the Indians remain unimpressed about what the United States accomplished. “We look at Pakistan's nuclear facilities harder than anyone else does, because we have the most to lose,” one high-ranking Indian general told me during a briefing on the country's military posture in New Delhi in April 2008. “Frankly, we have no confidence much has changed. Do you?”

In fact, at the end of Bush's term, the American officials who know the most about Pakistan's program are not as confident in private as they sound in public. For obvious reasons, Washington's official line is that the nuclear program is now secure. As long as the military—meaning Kidwai and the Strategic Plans Division—have ultimate control over the weapons, the constant chaos in the political leadership is unlikely to pose a threat, they insist. “It's a very professional military,” one senior American official told me during the worst of Pakistan's internal upheavals, in November 2007. “But the truth is, we don't know how many of the safeguards are institutionalized, and how many are dependent on Musharraf's guys.”²⁴

That fear of new leadership helps explain why Washington clung to Musharraf for so long, even to the point of appearing less than enthusiastic about the rise of a democratically elected government. “The nightmare scenario, of course, is what happens if an extremist Islamic government emerges—with an instant nuclear arsenal,” Robert Joseph, who helped design much of the American program to lock down the nuclear weapons, told me as Musharraf appeared on the edge of ouster in 2007.²⁵ But even after that fear passed—the Islamic parties did poorly in the elections in early 2008, even in the regions of the country where they have the strongest presence—there was plenty of reason to worry about what could happen in a time of political chaos.

In the 1990s, A. Q. Khan thrived in part because he knew that a Pakistani government distracted by its own leadership struggles would not be paying attention to his business dealings. His relationship with Iran flourished in the chaos that followed the death of President Muhammad Zia al-Haq in a plane crash in 1988. (Zia is buried outside a spectacular mosque that is just down the block from Khan's house.) The first deliveries of centrifuges to Iran took place as Benazir Bhutto was trying to secure her power. The deals with North Korea and Libya happened amid the political jockeying that brought Nawaz Sharif to office. A study of the Khan network published by the International Institute for Strategic Studies in London concluded that “the diffusion of domestic political power among the troika of the president, the prime minister, and the army chief obscured the command and control authority over the covert nuclear weapons program.”²⁶

Kidwai insists none of this could happen again; a decade after Pakistan's nuclear tests, the National Command Authority, he says, is far better established. He is almost certainly correct, though when I was in Pakistan in April 2008, interviewing members of the newly elected government, two top officials who were part of that command authority said they had no idea what it involved. “I haven't been briefed yet,” one confided to me. “I'm not sure what my role is supposed to be.” When I related this conversation to a top American official in Washington who asked me, on my return, what I had learned about the security of the Pakistani nuclear program, he grimaced and said, “Now you know what we've been dealing with for five years.”

In fact, despite their public enthusiasm about Kidwai's program, American officials cannot say precisely what has been accomplished. That is due partly to Pakistan's paranoia about not letting foreigners know where their weapons are, or exactly how they are protected. When a delegation of Americans went to Islamabad in the spring of 2007 to ask for more transparency concerning how American money was being spent, they were sent home empty-handed.²⁷

By any measure, the American aid devoted to locking up the world's most dangerous weapons in one of the world's most volatile countries is pocket change. The Pakistan program amounted to less than $100 million over five years, the equivalent of six hours or so of operations in Iraq in 2008. If Americans got to hold a referendum on which was more directly related to their security—ending the internal fighting in Baghdad or securing the weapons closest to al Qaeda's camps—it seems pretty clear how they would vote.

But as in most things in life, when it comes to preventing nuclear Armageddon, money isn't everything. There are, for example, nuclear rules—rules that are never posted, but that are supposed to be understood by all sides in a conflict. Even during the depths of the Cold War, there were rules in place that gave the Soviet premier and the American president some confidence that they knew how the other would act. Sometimes those rules stretched nearly to the breaking point, most notably during the Cuban Missile Crisis. As the Cold War dragged on, however, predictability became the greatest assurance of safety. We knew how far we could push the Soviets; they knew how far they could push us.

Only now are those rules beginning to be established between Pakistan and India. What is terrifying is that such rules will never exist between Pakistan and its newest enemy, the militant groups within Pakistan that see the country as their most promising target for takeover.

The absence of rules worried American officials during the tense months in 2001 and 2002 when it appeared, for a while, that the conflict between India and Pakistan could go nuclear. The incident that precipitated the hostilities was an attack on the Indian parliament in December 2001, just as the United States was consumed by the September 11 attacks and the invasion of Afghanistan. The attacks were the work of two terror groups within Pakistan, Jaish-e-Mohammed and Lashkar-e-Taiba, though there have always been suspicions that some elements of Musharraf's government may have been complicit. The crisis that followed led to the mass mobilization of the Indian Army, and a corresponding reaction in Pakistan. Colin Powell remembers scrambling to organize allies so that some foreign minister or world leader was visiting both New Delhi and Islamabad just about every week to talk both the Indian and Pakistani leadership down from their thinly veiled threats to use nuclear weapons. Musharraf was a particular worry because he was publicly warning that “we have means” to defend the country even if the Pakistani soldiers massing along the border were dwarfed by India's forces.

“We had sort of a duty roster out there for who is going tomorrow to keep these clowns from killing each other,” Powell told me in a conversation six years later. He recalled phoning Musharraf from Paris, with Bush sitting at his side, during a trip the president was taking through Russia and Europe. Powell's worry was that some hotheaded Pakistani field commander would take Musharraf's threats too literally.

“I called him from our embassy in Paris,” Powell recalled, warming to the tale. “I said, ‘Mr. President, how are you?’”

“‘Ahh, Mr. Secretary, how are you?’” Powell said, imitating Musharraf's distinctive accent.

“I said, ‘Well, I'm fine—umm, can we do general-to-general?’ And he said, ‘Of course, my friend.’ And I said all this talk about nuclear weapons or ‘means’ has got to stop. I said I was trained in this stuff and I know what it does and what it doesn't do and I also know that in this century no civilized person could ever think of taking such an existential step that has not been taken since 1945.

“‘It can't and won't happen,’ I said, but I told him, ‘General, you are scaring the crap out of everybody, so you've got to cool it.’ And he laughed and he said, ‘I understand.’”

Both sides stood back. They took a breath. They put their nuclear weapons back in their holsters—if they'd ever actually taken them out. (Kidwai argues that no weapons were moved during the crisis, though some American intelligence officials have their doubts.)

The very fact that Powell had to intervene, however, underscored the fact that Pakistan and India had never established “redlines.” Without them, the Pakistanis could not warn their commanders what kind of action might provoke a nuclear response by India. The Indians did not know what kind of action could trigger enough panic in Islamabad that they would roll out their arsenal.

It is worrisome because in the scenarios that American officials play out in private, militants might deliberately provoke another India-Pakistan crisis by launching a spectacular terror attack in India. At first it might seem as if the goal was to blow up the nascent peace process that has restored calm on the subcontinent. The real objective, however, may be darker: to provoke a conflict that would tempt Pakistan to move its nuclear weapons into place. That is when the arsenal would be most vulnerable, particularly if the Islamist groups are aided by insiders.

“Farfetched?” one senior American strategist said to me in 2008. “Maybe, but in Pakistan maybe not.”

When American officials “wargame” the next step, they ask the question, Would the Pakistani government tell Washington if one of its weapons was lost? American officials don't want to have to depend on Kidwai or his successor to pick up the hotline to deliver the bad news and to ask for a little help. For one thing, there's no hotline to pick up. For another, injured pride would likely trump prudence.

SO FAR, the American role in securing Pakistan's arsenal has been limited largely to training the people who, in turn, train the Pakistanis who operate or guard the country's weapons. American legalities—and paranoia on all sides—have prevented the United States from taking the next step: sharing the sophisticated electronic technology that for decades has successfully kept American weapons safe.^2*

The training has largely taken place at Sandia National Laboratories in New Mexico, where many of the government's nonprolifer-ation and nuclear-detection technologies are developed. Roughly 200 Pakistanis have learned the basics of protecting a nuclear arsenal at this facility. Most of this instruction was hardly rocket science. The lectures included how to build double fences, how to install security systems and motion sensors, and how to use night-vision goggles and radiation-detection devices. When it became clear that the Pakistanis had put a lot less effort into acquiring the right protective hardware than they had put into building a giant new reactor to produce plutonium, the United States paid for much of the safety equipment, shipping some of it directly to the Pakistani military. As soon as it was delivered, however, it disappeared into a black hole: The Pakistanis were not about to reveal where the warheads and missiles were stored.

Nonetheless, Hadley argues that even if you cannot verify the results, the American program is worth every penny. “I think it's exactly the kind of thing that the American people want us to be spending our money on,” he said one day in his office. “They just want to know the weapons are safe.” Now, with American support, Pakistan is supposed to be building a training center outside Islamabad to conduct the same training on Pakistani soil, but it is years delayed.

Hadley was less eager to discuss the administration's internal battle over whether to provide Pakistan with a far more advanced technology for securing weapons—something the nuclear experts call PALs, or “permissive action links,” a series of codes and hardware protections that make sure only a very small group of authorized users can arm and detonate a nuclear weapon.

PALs are a leftover from the Cold War, designed to make sure some rogue sergeant in a silo didn't wing a weapon toward Moscow. It may be more important in the Second Nuclear Age than it was in the first. When countries with little or no experience with nuclear weapons suddenly find themselves stacking them up in tunnels and caves, it would be nice to know that a terrorist who procured one could not simply set the timer and walk away.

In the American version, PALs hinge on what is essentially a switch in the firing circuit that requires the would-be user to enter a numeric code that starts a timer for the weapon's arming and detonation. If the sequence of numbers entered turns out to be incorrect in a fixed number of tries, the whole system disables itself. It is pretty similar to what happens when you repeatedly type the wrong passcode into an ATM machine, and the machine eats your bank card. But in this case, imagine that someone trying to use your stolen card entered the wrong code one time too many, and a series of small explosions was set off to wreck the innards of the bank machine. That's what happens to an American warhead—it is rendered useless. And in the American design, the system is buried deep inside the weapon so that no would-be terrorist could get inside to disable the safeguard system.²⁸

In an age of nuclear peril, it seems that this is precisely the kind of technology you would want to spread around the world to your friends, especially to friends who are better at building nukes than at figuring out how to keep them from going off. But nothing is that simple.

AFTER POWELL'S TRIP to Pakistan, a study provided to the White House concluded that giving PALs to the Pakistanis would violate both international and American law. After all, Pakistan was the ultimate nuclear outlaw: The country developed its arsenal covertly, never signed the Nuclear Non-Proliferation Treaty, and refused to allow investigators to question the greatest nuclear smuggler in history. Under U.S. law, Washington could not legally transfer nuclear technology to the Pakistanis, even if it was technology to make their weapons safer. Period. Ordinarily, one would not expect that to impede the Bush administration: In almost every other element of the war on terror—interrogating of suspects, setting up secret prisons, and inventing rules for military tribunals—the Bush administration rarely slowed down for legal niceties. But in this case they did.

To experts such as Harold M. Agnew, a former director of the Los Alamos weapons laboratory, where most of the United States’ nuclear arms were designed, the rules restricting the transfer of PALs are lunacy. They seem like a holdover of old-think, when one of the biggest fears was that we not teach other nations too much about how we secure our own weapons.

“Lawyers say it's classified,” Agnew told my Times colleague Bill Broad. “That's nonsense. We should share this technology. Anybody who joins the club should be helped to get this. Whether it's India or Pakistan or China or Iran, the most important thing is that you want to make sure there is no unauthorized use. You want to make sure that the guys who have their hands on the weapons can't use them without proper authorization.”²⁹

Unfortunately, the Pakistanis suspected that the United States had more in mind than just nuclear safety. Any PALs offered up in a FedEx box from Washington, they figured, would come with a secret “kill switch” that would allow someone deep inside the bowels of the Pentagon to track or disable Pakistan's nuclear assets. They were undoubtedly right.

Kidwai insists that he solved this problem by sending Pakistani engineers off to develop what one could only call “Pak-PALs,” an indigenous version of the American system. He told me that it was every bit as safe as the American version.

No one will talk about what role, if any, the United States played in designing this system. But history provides a possible clue. Back in the early 1970s, the United States faced a similar problem with a country that produces cheese the way Pakistan produces radicals: France. Worried about protecting the French arsenal, the United States began a series of highly secretive discussions with French scientists that amounted to a game of “twenty questions,” though in Washington-speak it was termed “negative guidance.” The process was detailed in a 1989 article in Foreign Policy by Richard Ullman.³⁰ According to that account, the French described their approach to building and securing a warhead, and the American nuclear scientists gave guidance about whether they were going off track. Just think of a blindfolded kid trying to pin the tail on a donkey, as everyone else shouts, “You're getting warmer.”

At this party, however, Washington was selective about who was allowed to play. When the race was on to secure Russia's arsenal, the United States hurried the declassification of limited information about American warheads so that it could be shared. China was another story: The Clinton administration determined that sharing PALs would be too risky, even though China was a signatory to all the right treaties. Officials feared giving the Chinese too many insights into how American systems worked, and it feared the backlash of seeming to sell another piece of critical technology to Beijing.

In the case of Pakistan, we may know how well Kidwai's “indigenous” PALs system works only after something happens. “Among the places in the world that we have to make sure we have done the maximum we can do, Pakistan is at the top of the list,” John McLaughlin, who served as deputy director of the Central Intelligence Agency during the investigation into A. Q. Khan, told me one afternoon in his small office at Johns Hopkins University. “I am confident of two things,” he added, “that the Pakistanis are very serious about securing this material, but also that someone in Pakistan is very intent on getting their hands on it.”

Naturally, no one in the U.S. government will say much of anything in public about what they know about Pakistan's arsenal, and even in background conversations top officials in the White House, the Pentagon, and the Energy Department are elusive. But, just as Kidwai fears, hardly a few months go by in Washington without someone conducting simulations of how the United States should respond if a terror group infiltrated the Pakistani nuclear program or managed to take control of one or two of its weapons. In these exercises, everyone plays to type: the State Department urges negotiations, while the Special Forces command loads its soldiers into airplanes. The results of these simulations are highly classified, for fear of tipping off the Pakistanis about what the United States knows or doesn't know about the location of the country's weapons. But as one frequent participant put it to me, “most of them don't end well.”

The problem with these exercises is that in the end the participants are never convinced that they could tell an American president, with confidence, that they know where all of Pakistan's weapons are—or that none are in the hands of Islamic extremists.

“It's worse than that,” a participant in one of the simulations told me. “We can't even certify exactly how many weapons the Pakistanis have—which makes it difficult to sound convincing that there's nothing to worry about.” Kidwai turned silent when I pressed him on the question of the size of Pakistan's arsenal. But it is clear that the number keeps changing because the country is upgrading its weaponry, building new, more efficient, plutonium-core weapons.

IN THE TOWER housing the International Atomic Energy Agency's offices in Vienna, the A. Q. Khan affair is not “closed,” despite Kid-wai's protestations. There are many unanswered questions remaining about his network, and one looming one. Who else possesses the design for Pakistan's bomb?

The question came up in earnest around 2006, when investigators in both the United States and Europe finally cracked the hard drives on Khan's computers, recovered from Bangkok, Dubai, and other locales where the Khan network was active.

The biggest trove was found in the computers that belonged to the Tinner family in Switzerland—the engineering specialists to whom the CIA had turned, and paid upward of $10 million, to provide information and help sabotage nuclear equipment going to Iran.

When Swiss police raided the Tinners’ offices and seized their computers, they found the Khan network's most important documents. There was a vast amount of material—orders for equipment, names and places where Khan's associates operated, even old love letters. There were several terabytes of data, a huge amount to sift through.

“There was stuff about dealing with Iranians in 2003, about how to avoid intelligence agents,” said one official who had reviewed it. “But then we got the new stuff—things that the local police had missed, maybe because it was encrypted.”

In 2008, announcing that the files had been destroyed by the Swiss government, Pascal Couchepin, the Swiss president, said they included “detailed construction plans for nuclear weapons, for gas ultracentrifuges for the enrichment of weapons-grade uranium, as well as for guided missile delivery systems.”

The most important of those plans was a digitized design for a nuclear bomb. But this was not the same design that the inspectors had seen before, the one the Libyans turned over on giant sheets of blueprint paper. Those sheets, while helpful to anyone building a weapon, “were more like an appetizer,” said the official who had seen both. This was the main course.

The plans were far more sophisticated than what had emerged from Libya. There were “timers and triggers in the design,” one investigator said. “Clearly someone had tried to modernize it, to improve the electronics. There were handwritten references to the electronics, and the question is, who was working on this?”

The design was not complete; it would have little value to a terrorist with no experience, but could have been quite helpful to a state, such as Iran, seeking to build an arsenal. The more that investigators examined the design, the more it became clear it was Pakistani.

It is unlikely that the work itself was done by A. Q. Khan. He was never a bomb designer. Instead it appeared to be a weapon design he had access to, and when the IAEA passed questions to Khan about whether this design had gone elsewhere, his response, one official told me, was “I cannot exclude the possibility.”^3*

The concern that raced through the upper echelons of the IAEA was that copies of this new design resided on the hard drives of several computers. It was impossible to know how many times it may have been copied.

The Swiss themselves were nervous about holding on to the bomb design—as became evident as a great drama played out between the Swiss and the CIA. The CIA did not want the Tinners prosecuted—not only would prosecution hurt the effort to recruit new nuclear spies, it would force the agency to acknowledge that the Tinners had worked for Washington. Moreover, at a moment when anti-Americanism in Pakistan was running high, the revelation that the bomb design was Pakistani would create even greater tension and undermine the effort to get the Pakistanis to cooperate. Eventually, the Swiss were persuaded to destroy all the material-even if it meant killing off any prosecution of the Tinners.³¹

Back in Washington, news of the new, more sophisticated design was closely held. One senior official I went to interview was happy to talk about the American effort to secure Pakistan's weapons. But when I started inquiring about the plans on the Khan computers, he hesitated. “I can confirm that they were there—we helped crack the code,” he said. “But beyond that, it's one of the few things we simply can't talk about.” When I asked about it at the White House and the upper echelons of the State and Defense Departments, some officials grimaced, others said they had never heard about it.

The design suggests that Khan was branching out, offering to sell not only the centrifuges to make bomb fuel but the blueprints to do something with it. “This looks to me like it was a cohesive plan,” one IAEA official said. When I asked Kidwai about it, he clearly knew about the discovery, but he waved away its importance. “What we've seen is incomplete,” he said. Anyway, he added, he did not believe that Khan would have had access to complete bomb designs, because they were done in a competing laboratory.

THE MAN WHO knows the answers to these questions is A. Q. Khan himself, and he isn't talking.

Except, of course, when the subject turns to regaining his freedom. Then he becomes garrulous. It didn't take long after the installation of a new government in Pakistan, in April 2008, for Khan to break his silence. With the rise in nationalistic fervor in the country—and the sense that it was time for the Americans to stop pushing Pakistan around—he saw a chance for his release.

“I saved the country for the first time when I made Pakistan a nuclear nation and saved it again when I confessed and took the whole blame on myself,” he told a reporter one evening in a telephone call.³² His family was pressing to allow him to have visitors and travel outside the house on Hillside Avenue, suggesting not so subtly that if the new government did not end his four-year-long home sentence, he might be tempted to talk about who else in the government was involved in his activities.

Appearing on television, the country's new foreign minister, Shah Mehmood Qureshi, said the time had come to restore Khan to a position of respect. “Yes, I don't want to see his movement restricted,” he told a Pakistani television station one day when I was reporting for this book in Islamabad. “He is a Pakistani, a respected Pakistani; I think that he should be allowed to see his friends, to go for a drive, to have a meal at a restaurant. I see no reason why he should be deprived of that; on the other hand we have to be concerned about his security and his health.”³³

In private, though, Khan was beginning to express to a few associates some second thoughts about his life's work. In January 2008, he wrote to one colleague that his desire to develop his nuclear expertise came from what he called the “traumatic” moments of his early life. He recalled traveling from Bhopal to Karachi with his family in 1947, amid horrific violence between Hindus and Muslims. He described the 1971 war in which Pakistan gave up “East Pakistan”—the war in which Kidwai was taken as a POW—which he termed a “disgraceful and humiliating surrender.”

“I got scared Indira Gandhi would go for the final kill,” he said. But in his old age, he had begun to discover that he empowered the army to rule over the country forever. “I never dreamt or thought of the disgraceful way the army would control the country under its boot,” he said. “They are here to stay on top of us for all time. I had thought that I was doing a patriotic service. Now it looks like it was a mistake.”³⁴

He ended the letter with a half-sentence: “No use of self-pity.”

1* By the time the paper published the story, many of the details had begun to leak out. In 2005, President Musharraf acknowledged in an on-the-record interview with us that he had received “international” help in securing the weapons, and the following year the Pakistani authorities gave a few more details to the local press in an effort to dispel rumors, which were never confirmed, that American personnel were involved in guarding Pakistani military sites. By the time of the 2007 instability in Pakistan, which appeared to threaten Musharraf's hold on power, the administration apparently decided that the secrecy around the program was impeding its ability to assure Americans that it had a plan in place to deal with the risks. “I think at this point,” one senior administration official said to me, “it's better that we make clear that we have attempted to help Pakistan through its security problems.”

2* During my visit to Pakistan in April 2008, Kidwai and other officials raised with me the issue of the series of cascading errors that led the U.S. Air Force to unknowingly fly a bunch of nuclear weapons on a B-52 bomber across the country in August 2007. There they were left at Barksdale Air Force Base in Louisiana for thirty-six hours with none of the special security required for nuclear devices, a point the Pakistanis enjoy making whenever the United States portrays them as vulnerable to “losing” a weapon.

3* These statements by Khan have to be treated with considerable skepticism. Questions aimed at Khan were forwarded by the IAEA and the CIA after he was put under house arrest in 2004. The Pakistani intelligence services say they brought the questions to him, and then delivered his answers back to Washington or Vienna. But neither agency had much understanding of how the questions were translated, or whether Khan's answers were edited by Pakistani authorities before they were returned to avoid statements that might contradict Pakistan's official line—that he had no access to weapons designs, and that no other Pakistani officials knew what he was doing.