Looking up, Dave saw another prisoner running toward him and firing a pistol from fewer than ten yards away. Dave shot him and then saw a large group, many still bound with rope, rushing toward him. Dave opened up with Spann's AK-47 while backpedaling. He later estimated that he shot at least fifteen before running out of ammunition and having to replace the empty magazine.

While running for cover, Dave stumbled over the bodies of several dead and wounded Uzbek guards. Eventually, he was able to reach temporary shelter in one of the buildings on the perimeter of the compound. There he ran into five foreign journalists, who asked his a.s.sistance in getting out of Qala-i-Jangi. Using one of the journalist's satellite phones, Dave called in reinforcements and air support. The small group holed up in various locations in the building for over five hours while a battle raged outside. During this period Dave was unsure about the status of his partner. One of the journalists said that he had seen Mike escape. As it started to get dark, Dave, the journalists, and several others managed to descend the north wall of the fortress and eventually reach safety.

It was a Sunday afternoon when I got word that we potentially had an officer down. I came into headquarters immediately to monitor developments. Shortly after 9/11, Cofer Black had told me that CIA might lose thirty to forty officers in carrying out our attack strategy. For a relatively small force such as ours, that was a stunning number. But even with such grim expectations-expectations that thankfully never were met-hearing that the first CIA officer was down struck us hard. I went to Hank Crumpton's small office in the CIA headquarters, where we waited in agony for hours, desperately trying to get information from the scene.

Despite the journalist's optimistic account of Mike Spann's escape, we feared the worst for him. Two painful days would pa.s.s before U.S. and Afghan allied forces could put down the rebellion, get inside the fortress, and determine for certain that Mike was dead. Word of the riot and the possible death of a U.S. official did not wait for confirmation. Reports of the clash were soon airing around the world, and Pentagon spokesmen were quick to tell the media that no U.S. military personnel were unaccounted for. That led reporters to leap quickly and accurately to the conclusion that a CIA officer was the victim.

Mike Spann was a thirty-two-year-old former Marine who had been with CIA for only a short period. His wife, Shannon, was also a member of the CIA's clandestine service and was on the West Coast with her infant son, visiting family, at the time of the attack. Shannon was out driving when she heard a radio report about the possibility of a CIA officer missing. Immediately, she pulled her car over to the side of the road and called headquarters to find out what she could. I dispatched some officers to California to be with her, and others to Alabama to a.s.sist Mike's parents, even before we were able to verify his status.

Once Mike's body was recovered and his family informed, we made the decision to confirm his death to the media. Such confirmation is routine for the military but not always so for CIA. In this case, however, the fact of Mike's Agency background had already leaked. His family wanted to acknowledge who he was and express their pride in his service. There was no way to keep the Agency connection secret and little reason to try. Yet we were quickly criticized by pundits, who accused us of seeking publicity over the first American to die in combat in Afghanistan.

As it turned out, I had to take a trip to Pakistan shortly after Mike was killed to meet with President Musharraf over urgent intelligence we had received regarding possible follow-up al-Qa'ida attacks against the United States. On the way back to the United States, I had my plane divert to Germany, where Mike's body had been taken. On December 2, we brought him on his final trip home. I've never made a more somber journey.

Eight days later, Mike Spann was buried with full honors at Arlington National Cemetery. Shannon impressed us all with her grace, dignity, and strength. The family asked me to make remarks at the graveside, and I was honored to do so. In going to Afghanistan, to "that place of danger and terror, he sought to bring justice and freedom," I said. I told his family, friends, colleagues, and the nation that Mike Spann was a "patriot who knew that information saves lives, and that its collection is a risk worth taking."

The bureaucracy had initially balked at burying Mike in Arlington, since he had been neither retired military nor on active duty at the time of his death. John McLaughlin called Paul Wolfowitz, who quickly said he would support Mike's being given the honor of an Arlington interment. John then called Andy Card, who, based on McLaughlin and Wolfowitz's recommendation, cut through the red tape and made it happen.

Mike's is one of many remarkable stories of heroism by CIA officers in the opening months of the Afghan campaign. Although they are accustomed to working without much support or infrastructure, Afghanistan took that to new heights. Agency officers partic.i.p.ated in cavalry charges and called in air strikes while on horseback. One CIA medic attempted to save an Afghan's life by performing an emergency amputation of the soldier's leg using the only device available to him-a large Leatherman pocketknife.

The definitive moment of the CIA's entire campaign may have been saving the life of the country's future leader. By the very early days of December, Hamid Karzai had proved himself not only a fearless fighter but also the indispensable man in the Afghan equation. As a result, CIA and U.S. Special Forces units began to worry about not just supporting him but also ensuring his survival. That, though, became increasingly difficult.

On December 5, Karzai was leading his troops in an a.s.sault on Khandahar, one of the last Taliban strongholds. U.S. military personnel were calling in air strikes in support of the a.s.sault using Global Positioning System devices. As they were doing so, one soldier replaced the batteries in his GPS unit, forgetting that doing so caused the unit to erase previously entered data and to reset itself at its own location. As a result, an air strike from a B-52 was called in on the soldier's own position. Three Americans and five Afghans died in the mishap. Karzai might have, too, if Greg V. hadn't thrown himself on him, knocking him to the ground just as the bombs struck. It turned out to be an eventful Wednesday for Karzai. That same day, he was selected to be the interim prime minister of Afghanistan.

The routing of the Taliban and al-Qa'ida from Afghanistan in a matter of weeks was accomplished by 110 CIA officers, 316 Special Forces personnel, and scores of Joint Special Operations Command raiders creating havoc behind enemy lines-a band of brothers with the support of U.S. airpower, following a CIA plan, that has to rank as one of the great successes in Agency history.

As we forced out al-Qa'ida and the Taliban leaders from the sanctuary, we continued our focus on capturing or killing Usama bin Ladin. We believed he was in the mountains of southern Nangarhar province, only miles from the Pakistani border. This area had long been an al-Qa'ida stronghold, particularly south of Jalalabad in the Tora Bora Mountains.

By early November, our intelligence reporting was indicating that UBL had fled to the Tora Bora region. When Kabul fell, on November 14, we figured that Bin Ladin and his cohorts would be even more likely to try to flee Afghanistan, perhaps for the ungoverned regions of Pakistan. CIA rushed to set up counterterrorist pursuit teams, made up of Northern Alliance fighters with U.S. advisors, but the vast reaches of the territory made this a difficult mission. Bin Ladin had chosen a good place to hide. The rugged hills of Tora Bora contain dozens of tunnels and caves. As one CIA officer put it, "He had mountains to his back, clear fields of fire in front of him, and a local population unwilling to confront or eject him."

Agency and military officers tried to motivate Afghan forces with the usual combination of exhortations and a liberal allocation of cash to press the attack against suspected al-Qa'ida strongholds. A joint CIA/JSOC team of five men infiltrated into the heart of enemy territory and, for more than seventy-two hours, directed air strikes. At one point, the team requested B-52s to deliver bombs to within twelve hundred yards of their position. In all, about seven hundred thousand pounds of ordnance was dropped between the fourth and seventh of December alone. Hundreds of al-Qa'ida operatives were killed. But CIA officers on the scene began to doubt whether they could rely on the Afghan ground force for this critical push of the campaign. Worse, there were concerns that some of the Afghan units might be actively cooperating with al-Qa'ida elements, helping them to escape.

We had sensitive intelligence that strongly suggested Bin Ladin was in the Tora Bora area and likely was plotting a quick escape through soon-to-be-completed tunnels. U.S. air power was brought to bear on this very difficult terrain.

Aerial bombardment, though, can do only so much. Truly confronting an enemy entrenched in a network of caves requires getting into the caves yourself, and the Afghan troops we were working with were distinctly reluctant to undertake that risk. It was also the Islamic holy month of Ramadan, and Afghan troops were not much interested in conducting attacks. Agency officers in the field and at headquarters started lobbying hard for the insertion of U.S. troops to try to complete the job. Hank Crumpton called Tommy Franks to discuss the situation. Tommy said that if he were to deploy a large contingent of U.S. military to the region it would take weeks to get them in place during that period, and UBL might slip away. He made the call that it was better to press ahead with the units in place at the moment than to wait for reinforcements. We urged the Pakistanis to do the best they could to place troops along the Pakistani-Afghan border. We plotted all available escape routes that Bin Ladin might choose.

I remember the president asking Hank one morning if the Pakistanis could seal the border. "No sir," he said. "No one has enough troops to prevent any possibility of escape in a region like that." The Pakistani military did manage to capture hundreds of al-Qa'ida members slipping across the border, but not the one we wanted most.

CHAPTER 13

Threat Matrix

The attacks of 9/11 were not the end of anything. They were the beginning. That was the message I was getting from my Counterterrorism Center. As far as al-Qa'ida was concerned, 9/11 was just the opening shot.

As traumatic as the attacks were, however, we knew what actions we could take. We knew what needed to be done, and there was a tremendous sense of urgency about it. Over the next several years we were able to achieve remarkable success against the terrorist threat for three strategic reasons.

First was the loss of al-Qa'ida's safe haven in Afghanistan. Because we were able to get into the sanctuary, we suddenly had access to people and doc.u.ments that laid bare the future plans and intentions of al-Qa'ida. The key to success was rapidly to collect, fuse, and a.n.a.lyze the data in real time and to use it to drive operations.

The second strategic reason for success was Pakistani president Musharraf's decision to join the fight on our side. Pakistan switched sides-from aiding the Taliban to fighting al-Qa'ida. Pakistani intelligence chief Ehsan Ulhaq became a pivotal figure. With the arrest of well over five hundred al-Qa'ida operatives, Pakistan, in concert with U.S. intelligence, denied al-Qa'ida the luxury of a safe haven within the country's settled areas. (For his efforts, al-Qa'ida twice tried to a.s.sa.s.sinate President Musharraf.) The third reason was the decisive action on the part of the Saudi leadership following the Riyadh bombings in May 2003. Saudi authorities have detained or killed many of the top known al-Qa'ida cell leaders in the kingdom and hundreds of foot soldiers. They have captured thousands of pounds of explosives. They have also reduced the financial resources at al-Qa'ida's disposal.

Afghanistan, Pakistan, and Saudi Arabia were just part of the puzzle. With the new authorities, money, and confidence that the U.S. president gave to us, we were able to leverage the rest of the world's counterterrorism efforts.

There were a few countries that "got it" long before 9/11. The Jordanians, Egyptians, Uzbeks, Moroccans, and Algerians always understood what we were talking about. It was ironic that, pre-9/11, we had more success in getting help within the Islamic world than elsewhere. The British and French were also always helpful. Both had lived through their own terrorist threats. But until September 11, it was hard to convince most of the world of the legitimacy of our concerns.

In addition to the strategic reasons for our success, there were several tactical steps that were important. One of the most significant keys to our accomplishments against the terrorists came from something that sounds quite mundane: a daily meeting. This meeting would be repeated at 5:00 P.M. P.M. every weekday for the three years after 9/11. At these sessions we would try to get a handle on the flood of information about terrorism pouring in from around the world. Virtually every day you would hear something about a possible impending threat that would scare you to death. But you would also hear about opportunities to work with allies, new and old, against this threat. These sessions grew out of biweekly terrorism update meetings I started when I was deputy DCI in 1996. In 1998, after the emba.s.sy bombings, the meetings became weekly. Initially we called it "the small group." That t.i.tle quickly became a joke, because the number of partic.i.p.ants expanded until they packed the large wood-paneled conference room down the hall from my office. every weekday for the three years after 9/11. At these sessions we would try to get a handle on the flood of information about terrorism pouring in from around the world. Virtually every day you would hear something about a possible impending threat that would scare you to death. But you would also hear about opportunities to work with allies, new and old, against this threat. These sessions grew out of biweekly terrorism update meetings I started when I was deputy DCI in 1996. In 1998, after the emba.s.sy bombings, the meetings became weekly. Initially we called it "the small group." That t.i.tle quickly became a joke, because the number of partic.i.p.ants expanded until they packed the large wood-paneled conference room down the hall from my office.

The point of the meeting was to pull together in one place everyone who needed to take action in the next twenty-four hours in both our war in Afghanistan and the broader war on terrorism. My intent was to cut short the time it took for information to flow from the people in the field to me and to slash the time between orders being issued in Washington and executed half a world away.

This wasn't CIA talking to itself; we had FBI, NSA, and military officers there as well. The windowless room features a long, highly polished wooden conference table with about twenty chairs around it. The conference room needed its long table because briefers would occasionally roll out charts the size of bedsheets showing a.n.a.lysis that connected terrorists around the world through family, phone, and/or financial contacts. Just before the session started, any maps, charts, or doc.u.ments to be used in the presentations would be pa.s.sed out, and at the end they would be just as efficiently collected to keep control of the information. Always there was a palpable fear in the room that the United States was about to be hit again-either here or our interests abroad. No one present thought there was a minute to waste.

Five or six Agency components would lead off the meeting every afternoon. The first briefer was usually from the Office of Terrorism a.n.a.lysis, initially Pattie Kindsvater, Phil Mudd, and other a.n.a.lysts. Later it was Mark Rosini from the FBI, whom we affectionately called "The Voice," because his deep baritone imparted a special sense of urgency. These briefers would run down the latest threat information. The terrorist acts of 9/11 unleashed a torrent of information from around the world. Suddenly friend and foe alike started reporting information that a day or two earlier they might have withheld or ignored. Some of it would later prove to be questionable, but at the time, we could not afford to dismiss any potential threat-and there were thousands of them.

To help senior administration officials visualize the range of possible plots we were tracking, we developed, in coordination with the FBI, what we called the "threat matrix." A multipage doc.u.ment, the matrix was given to the president each morning as part of his PDB session. Copies of it were also provided to other top officials. In it were the newest threats that had emerged over the past twenty-four hours.

The matrix soon became an important part of the five o'clock meeting. At each session, we went over the next day's matrix, recognizing that many, perhaps most, of the threats contained in it were bogus. We just didn't know which ones. In a typical matrix you might see tales of impending doom picked up from people walking into U.S. emba.s.sies overseas, cryptic comments gathered through intercepted foreign communications, anonymous correspondence received by major media outlets, and leads given to us by human a.s.sets.

We recognized that the matrix was a blunt instrument. You could drive yourself crazy believing all or even half of what was in it. It was exceptionally useful, however, and an unprecedented mechanism for systematically organizing, tracking, validating, cross-checking, and debunking the voluminous amount of threat data flowing into the intelligence community. The very ma.s.siveness of it prompted officials to think through vulnerabilities. Have we done enough to secure major landmarks, theme parks, or water supplies? Are our watchlists tight enough? Sometimes the threats mentioned would strike you as absurd, and then al-Qa'ida would do something to convince you that nothing was out of the range of possibility. Who, for example, would have thought that exploding footwear could be a major air travel problem-until, that is, December 21, 2001, when Richard Reid was subdued on an American Airlines flight from Paris to Miami trying to light explosives hidden in his shoes?

After the discussion of the threat matrix, Hank Crumpton, chief of the CTC Special Operations Group, would come next. He'd be followed by the chief of Alec Station's Bin Ladin Unit, initially Hendrik V., and later Marty M.; then Rolf Mowatt-Larssen, head of CTC's WMD branch, would brief. On occasion we would hear from Phil R., who was in charge of CTC's efforts involving international financial operations. Charlie Allen would carefully listen to our operational requirements and translate them into information requirements, which our intelligence communities, both foreign and domestic, would have to pursue. This was both to meet imminent operational needs and to position us to stay one step ahead of the terrorists.

Also at my side at the five o'clock meetings were John McLaughlin; the heads of the Directorates of Operations, Intelligence, and Science and Technology; the senior leadership of CTC; and others whose goal was to help clear obstacles for those who were on the front lines. Attendance at the five o'clock meetings became a critical part of each person's day. If, for some reason, you missed a meeting, you'd have to struggle the next day to follow the plot lines-so much interconnected information flowed each time.

November 6, 2001, was a typical five o'clock session. On that day I was briefed on a wide variety of freshly collected intelligence: A report had been collected about an Arab, of Persian Gulf origin, who reportedly knew of a planned second strike against the United States that was imminent and who claimed that the operatives were already in place. Additionally, he claimed to know of a third and final attack after which he would be free to come home. Similarly there was information on someone apparently in Jordan who had posted on a website a prediction that another attack on the United States was imminent. You might ask, so what? Until you learned that this same person had posted a note saying they were close to "zero hour" on September 10, 2001.

Another snippet of intelligence that day told us that a known al-Qa'ida a.s.sociate who had been in the United States from 1999 to the fall of 2001 was aware of big events expected on November 5 and 6. We also learned that an Egyptian who worked for the emba.s.sy in Saudi Arabia had suddenly, without explanation, faxed in his resignation. Subsequent investigation showed that the man had ties to al-Qa'ida's partner, Egyptian Islamic Jihad, and was wanted by authorities in his home country. Could his disappearance presage some new attack? We had to try to find him fast.

That same evening, I heard about intelligence gleaned from a senior UBL operative that provided the name of an al-Qa'ida a.s.sociate determined to conduct a suicide operation. We had the name, biographical data, but no idea where the man was.

Nearly two months after the attacks of 9/11 there was still great skepticism in Saudi Arabia that any of their countrymen had been involved. My staff came to me that night with a proposal that we share the chilling c.o.c.kpit audio recordings made from United Airlines Flight 93 before it crashed in Pennsylvania. The Saudi-accented voices heard on the tape might remove any doubts.

We had intelligence of three al-Qa'idaa.s.sociated people, possibly connected to Abu Musab al-Zarqawi, traveling for unknown reasons; we pa.s.sed along the intelligence to three countries, all mentioned as possible transit points.

We heard from Russian intelligence about increased concerns over terrorist actions in Chechnya.

A Middle Eastern country captured a terrorist wanted in a third country. Could we help get him there? We could.

The FBI had conducted a polygraph on a source of the U.S. Customs Service who said he knew of a possible nuclear threat to the United States; that source flunked the test, which showed "deception indicated."

The intelligence we heard that night, and every night, were just tiny threads. They had to be woven into a tapestry before we could make sense of what we were seeing. And this was just one day; it is difficult to put in words the number of reports, and the intensity of those reports, that came in every day. As one officer said to me, "I never want to live that again. The pace was furious. The constant refrain was: It must be done tonight, it must be done tomorrow. We have to have that for the president tomorrow. That pace wasn't kept up for days or weeks; it was years."

The five o'clock meetings were decision-making sessions, not briefings. If someone told me he was having trouble getting needed information out of an allied government, I'd often grab the phone right after leaving the meeting, call the head of the intelligence service involved, and light a fire under him. Other times I would order up talking points to be in my hands by six the next morning.

Other governments weren't the only concern. Sometimes we would hear of potential threats that weren't being internalized quickly enough within our own own government. Countless times someone in the room was directed to get up that second, find a phone, and call the Pentagon, the FBI, the State Department, or some other ent.i.ty, to make absolutely sure that the right people knew everything we knew and that they were going to get on top of that particular threat. The key was imparting information and context quickly; we had no time for more briefings. government. Countless times someone in the room was directed to get up that second, find a phone, and call the Pentagon, the FBI, the State Department, or some other ent.i.ty, to make absolutely sure that the right people knew everything we knew and that they were going to get on top of that particular threat. The key was imparting information and context quickly; we had no time for more briefings.

On many occasions, I would be briefed on matters that were, as they say in Washington, "outside my lane." When that happened, I would tell, say, the FBI representative to call Director Bob Mueller and bring him up to speed on a domestic issue, because we intended to mention it in the next day's PDB session in the Oval Office. Without doubt, the president was going to turn to Bob and ask what he was doing about this; it was in everyone's interest that he had a good answer.

Our morning sessions with the president were also intense. He quickly became steeped in our strategy, with regard to activities not only in Afghanistan but also in the rest of the world. He was focused on results yet at the same time did not seek to micromanage our operations. He spent time with the substantive experts we brought to daily meetings and to longer sessions at Camp David on Sat.u.r.days. The president never became the action officer, but there was no doubt the leader was in the trenches with us. If you told him about an imminent operation on Monday, you could be certain after a few days he would ask about it, if we had not provided the necessary follow-up.

A PDB session would lead to a broader meeting with Bob Mueller, Tom Ridge, later Fran Townsend, and their staffs, to review the threat matrix, the actions that were being taken, the gaps in our knowledge, and the interventions the president or vice president could undertake to help. Over time, at Andy Card's insistence, we modified the items in the matrix the president would see, to ensure that only those with the necessary weight and quality consumed his attention. When you have been accused of failing to connect the dots, your initial reaction is to ensure that all the dots are briefed. Until our knowledge became more refined, our inclination was to overbrief.

At the core of our effort was the Counterterrorism Center. It was the hub around which all of our efforts revolved. From there CIA stations worldwide were tapped to work both unilaterally and with host government intelligence services to improve the information sharing we relied upon. The long-standing relationships that Agency officers had with counterparts around the world became essential to our success. Even former adversaries seemed more willing to work with us.

As we made progress overseas, we found ourselves struggling domestically. It was stunning how little reliable information was immediately available inside our own borders. There was no good data on how many foreigners had overstayed their visas and no tracking system to see if young men who came into this country to attend university had actually shown up for cla.s.ses-or if they had changed their major from music to nuclear physics. Nor was there any way for a police department in one part of the country to share suspicious activity data with counterparts across the state or the nation. There was no seamless way to communicate from Beirut to Seattle; there was no communications backbone. And while there were mountains of data within the United States, no one knew how to access it all, and little had been done to train people to put it together and report it, much less a.n.a.lyze it. In the early days, what we did not know about what was going on in the United States haunted us. We had to make judgments based on instinct.

Few understand the palpable sense of uncertainty and even fear that gripped those in the storm's center in the immediate aftermath of 9/11. One particular concern was the fact that, although there wasn't any tracking system in place, there were thousands of foreigners in the United States whose visas had expired. The most important thing we needed to do was to prove the negative: that there were not more al-Qa'ida cells within the country poised to conduct a second wave of attacks. At the time, I remember reflecting on testimony Gen. Mike Hayden, then the director of NSA, had given to a public hearing of the House Intelligence Committee in 2000. Mike created quite a stir when he said that if Usama bin Ladin had crossed the bridge from Niagara Falls, Ontario, to Niagara Falls, New York, there were provisions of U.S. law that would offer him protections with regard to how NSA could cover him. Mike would later say that he was using this as a stark hypothetical. On September 12, 2001, it became real.

After the 9/11 attacks, using his existing authorities, Hayden implemented a program to monitor communications to and from Afghanistan, where the 9/11 attacks were planned. With regard to NSA's policy of minimization, balancing U.S. privacy and inherent intelligence value, Mike moved from a peacetime to a wartime standard. He briefed me on this, and I approved. By early October 2001, Hayden had briefed the full House Intelligence Committee and the leadership of the Senate Intelligence Committee.

Soon thereafter, the vice president asked me if NSA could do more. Our ability to monitor al-Qa'ida's planning was limited because of constraints we had imposed on ourselves through the pa.s.sing of certain U.S. laws in the late 1970s. I called Mike to relay the vice president's inquiry. Mike made it clear that he could do no more within the existing authorities. We went to see the vice president together. Mike laid out what could be done that would be feasible, prudent, and effective.

Within a week new authorities were granted to allow NSA to pursue what is now known as the "terrorist surveillance program." The rules required that at least one side of the phone call being surveilled be outside the United States and that there be probable cause to believe that at least one end of the communication was with someone a.s.sociated with al-Qa'ida. Elaborate protocols were set up to ensure that the program was carried out in accordance with these regulations. Within weeks of the program's inception, senior congressional leaders were called to the White House and briefed on it. Prior to its disclosure, twelve such briefings were hosted by the vice president for the leaders of the House and Senate Intelligence Committees. The briefings were thorough and disciplined. From my perspective, Mike gave the members full insight into how the program was being managed, the care that was being taken to ensure that it lived up to its intent, and offered the best a.n.a.lysis he could provide with regard to its results. The program was reauthorized by the president about every forty-five days prior to its disclosure. Each reauthorization was accompanied by an intelligence review, each of which I signed prior to my retirement. This included a comprehensive a.s.sessment of the value of continuing the program.

At one point in 2004 there was even a discussion with the congressional leadership in the White House Situation Room with regard to whether new legislation should be introduced to amend the FISA statute, to put the program on a broader legal foundation. The view that day on the part of members of Congress was that this could not be done without jeopardizing the program.

Mike Hayden has persuasively argued that the FISA statute enacted in 1978 could not have contemplated the technology available for terrorist use today, nor provided for the speed needed to deter today's terrorist acts. A bipartisan effort to amend the statute would be wise, so long as it is done in a manner that does not jeopardize critical operational equities. The trauma of 9/11 led, in the words of Mike Hayden, to a program to protect our liberty by making us all feel safer. It was never about violating the privacy of our citizens.

Had this program existed prior to 9/11, Mike Hayden has said that, in his professional judgment, we would have detected some of the al-Qa'ida operatives in the United States and we would have identified them as such. I agree.

As we were coming up with the new terrorist surveillance program, our working a.s.sumption had always been that the attacks of 9/11 were simply the first wave. Al-Qa'ida had declared its intention to destroy our country. Why then would it be satisfied with just three thousand deaths? It was inconceivable to us that Bin Ladin had not already positioned people to conduct second, and possibly third and fourth waves of attacks inside the United States. Getting people into this country-legally or illegally-was no challenge before 9/11. Al-Qa'ida had to have known that things would tighten up after the attacks, so logic suggested that they would have acted in advance to prepare for that inevitability. We considered the possibility that in addition to carrying out the September 11 attacks, the nineteen hijackers might also have done casing and provided surveillance for whatever attack would come next. Nothing that I learned in the ensuing three years ever led me to believe that our initial working a.s.sumption that al-Qa'ida had cells here was wrong.

Increasingly, we began to concentrate on the possible connections between the domestic front and the data we were collecting overseas. We would identify al-Qa'ida members and other terrorists overseas and often discover that they had relatives, acquaintances, or business ties with people in the United States. Each rock overturned abroad led to ants scurrying every which way, including many toward the United States. These concerns, in part, led to the establishment of the NSA program wrongly described by the media as "domestic spying." The program grew out of concrete evidence that foreign terrorists planning new attacks on America were in communication with colleagues in this country. Oddly, the farther terrorists were from our sh.o.r.es, the more vulnerable they were to our intelligence-collection efforts. In some ways, the safest place for an al-Qa'ida member to hide was inside the United States.

As much as our government would have liked to capture or kill Usama bin Ladin and Ayman al-Zawahiri, we recognized that the key to crippling al-Qa'ida would be to take down the next tier of leadership, the facilitators, planners, financiers, doc.u.ment forgers, and the like. These were the people who would have the actual links to the terrorist operatives. If we could disrupt or destroy the efforts of these individuals, we might prevent the follow-on attack that we feared so much. Our strategy was clear: to weaken al-Qa'ida's ability to plan and execute attacks, by forcing them to move less capable individuals into positions of leadership. In particular, our focus was on the individuals in charge of planning operations against the United States. Once Khalid Sheikh Mohammed was captured, Abu Faraj al-Libi took over. He was captured in Pakistan in May 2005 and replaced by Hamza Rabi'a, who was reportedly killed in the North Waziristan province of Pakistan seven months later.

One of the first dominoes to fall was Abu Zubaydah. Before 9/11, his name had been all over our threat reporting. After the attacks, he gained an even more prominent role in al-Qa'ida, especially once the United States killed the group's number three man, Mohammed Atef, in a November 2001 air strike in Afghanistan. Time and again in our five o'clock meeting we discussed how to run Abu Zubaydah to the ground.

By March 2002 we had identified a large number of sites in Pakistan that appeared to be al-Qa'ida safe houses. We got the increasingly helpful Pakistani authorities to raid thirteen of them simultaneously; they captured more than two dozen al-Qa'ida members. We were hopeful that a big fish like Abu Zubaydah would be in one of the safe houses, and we were not disappointed. In Pakistan's third largest city, Faisalabad, a gunfight broke out when Pakistani security officials stormed a second-floor apartment. Abu Zubaydah, who was inside, was shot three times and critically wounded.

Ironically, we found ourselves suddenly concerned with trying to save a terrorist's life. Not that we had any sympathy for Zubaydah; we just didn't want him dying before we could learn what he might have to tell us about plans for future attacks. Fortunately, Buzzy Krongard, our executive director, was also on the board of directors of Johns Hopkins Medical Center. Using his contacts there, he arranged for a world-cla.s.s medical expert to jump aboard an aircraft we had chartered so he could be flown to Pakistan and save a killer's life. Once Abu Zubaydah was stabilized, the Pakistanis turned him over to CIA custody. It was at this point that we got into holding and interrogating high-value detainees-"HVDs," as we called them-in a serious way.

Detainees, in general, had become a critical issue. By this time, many Taliban and al-Qa'ida prisoners were in military custody. Yet the quant.i.ty and quality of intelligence produced from their interrogation was disappointing. The detainees were either too low ranking to know much or too disciplined to reveal useful information.

Abu Zubaydah's capture altered that equation. Now that we had an undoubted resource in our hands-the highest-ranking al-Qa'ida official captured to date-we opened discussions within the National Security Council as to how to handle him, since holding and interrogating large numbers of al-Qa'ida operatives had never been part of our plan. But Zubaydah and a small number of other extremely highly placed terrorists potentially had information that might save thousands of lives. We wondered what we could legitimately do to get that information. Despite what Hollywood might have you believe, in situations like this you don't call in the tough guys; you call in the lawyers. It took until August to get clear guidance on what Agency officers could legally do. Without such legal determinations from the Department of Justice, our officers would have been at risk for future second-guessing. We knew that, like almost everything else in Washington, the program would eventually be leaked and our Agency and its people would be inaccurately portrayed in the worst possible light. Out of those conversations came a decision that CIA would hold and interrogate a small number of HVDs.

CIA officers came up with a series of interrogation techniques that would be carefully monitored at all times to ensure the safety of the prisoner. The administration and the Department of Justice were fully briefed and approved the use of these tactics. After we received written Department of Justice guidance on the interrogation issue, we briefed the chairmen and ranking members of our oversight committees. While they were not asked to formally approve the program, as it was conducted under the president's unilateral authorities, I can recall no objections being raised.

The most aggressive interrogation techniques conducted by CIA personnel were applied to only a handful of the worst terrorists on the planet, including people who had planned the 9/11 attacks and who, among other things, were responsible for journalist Daniel Pearl's death. The interrogation of these few individuals was conducted in a precisely monitored, measured way intended to try to prevent what we believed to be an imminent follow-on attack. Information from these interrogations helped disrupt plots aimed at locations in the United States, the United Kingdom, the Middle East, South Asia, and Central Asia.

The president confirmed the existence of the interrogation program on September 6, 2006, when he announced that fourteen HVDs who had been held under CIA control would be transferred to Guantanamo Bay.

Like many of the al-Qa'ida detainees, Abu Zubaydah originally thought that he could outsmart his questioners. He would offer up bits and pieces of information that he thought would give the impression of his providing useful material, without really compromising operational security.

But Abu Zubaydah ultimately provided a motherlode of information, and not just from his interrogation. We were able to exploit data found on his cell phone, computer, and doc.u.ments in his possession that greatly added to our understanding of his contacts and involvement in terrorism plotting.

Interrogating Abu Zubaydah led us to Ramzi bin al-Shibh. A Yemeni by birth, Bin al-Shibh had studied in Germany with three of the eventual 9/11 hijackers. He had intended to be one of them and was deterred only after four attempts to obtain a U.S. visa failed. Instead, he served as the primary communication link between the hijackers and al-Qa'ida central, meeting with the plot's ringleader, Mohammed Atta, in Germany and Spain, and staying in touch with the terrorists via phone and e-mail. With Zubaydah's unintentional help, Bin al-Shibh was captured by Pakistani authorities on the first anniversary of the 9/11 attacks, after a gun battle in Karachi.

But no success story lasts long in Washington before someone tries to minimize it. A published report in 2006 contended that Abu Zubaydah was mentally unstable and that the administration had overstated his importance. Baloney. Abu Zubaydah had been at the crossroads of many al-Qa'ida operations and was in position to-and did-share critical information with his interrogators. Apparently, the source of the rumor that Abu Zubaydah was unbalanced was his personal diary, in which he adopted various personas. From that shaky perch, some junior Freudians leapt to the conclusion that Zubaydah had multiple personalities. In fact, Agency psychiatrists eventually determined that in his diary he was using a sophisticated literary device to express himself. And, boy, did he express himself.

Abu Zubaydah's diary was hundreds of pages long. Agency linguists translated enough of it to determine there was nothing of operational use in it, yet some Pentagon officials, including Paul Wolfowitz, seemed fascinated with the subject and kept bugging us to translate the whole doc.u.ment. We kept resisting. One day Wolfowitz hounded his CIA briefer. "Why wouldn't we devote the resources to convert the book to English?" he demanded. "We know enough about the diary," the briefer explained, "to know that it simply contains a young man's thoughts about life-and especially about what he wanted to do with women." "Well, what have you learned from that?" Wolfowitz asked. Without missing a beat, the briefer responded, "That men are pigs!" Wolfowitz's military a.s.sistant laughed so hard he fell off his chair.

But in Afghanistan there was no time for laughter. As we achieved success in driving al-Qa'ida out of Afghanistan, they began to search for other sanctuaries for their leadership. The organization sought places where they could plan future attacks against the United States with impunity from law enforcement, intelligence, and military operations. First, al-Qa'ida established itself in the settled areas of Pakistan. Later they moved into the ungoverned tribal areas of South Waziristan. Later still, Pakistani military operations drove them farther north, to areas where I believe their senior leaders continue to operate.

In mid-2002 we learned that portions of al-Qa'ida's leadership structure had relocated to Iran. This became much more problematic, leading to overtures to Iran and eventually face-to-face discussions with Iranian officials in December 2002 and early 2003. Ultimately, the al-Qa'ida leaders in Iran were placed under some form of house arrest, although the Iranians refused to deport them to their countries of origin, as we had requested.

In the spring of 2002, computers, phone records, and other data from al-Qa'ida takedowns in Pakistan, Afghanistan, and elsewhere started to suggest troubling connections to individuals in the United States, particularly in the Buffalo, New York, area. As with so much else in those hectic days, I first learned about all this at one of our five o'clock meetings. I told the lead a.n.a.lyst on the matter to share her concerns immediately with the FBI. We had her take all her data to the regional FBI office, where, initially, she got a skeptical reception. Even in the aftermath of 9/11, there was a reluctance to believe that sleeper cells could be operating in the United States, particularly cells made up of American citizens. But as the FBI dug into the matter, the Bureau became believers. Six Yemeni Americans, all of whom had received training at an al-Qa'ida camp in Afghanistan prior to 9/11, were arrested in September 2002. The group, which became known as the Lackawanna Six, later pled guilty to terrorism-related charges and received prison terms ranging from eight to ten years each.

The five o'clock meetings did more than coordinate the takedown of individual terrorists and unravel future plots. We also used them to track the ebb and flow of overall threat concerns. Throughout the three years after 9/11 there was a lot more "flow" than there was "ebb."

These times of heightened concern would often translate into increasing the terrorist threat warning levels from yellow to orange. We did so on four occasions. In each instance there was a credible intelligence basis for doing so. Initially, there was no choice but to burden the entire country. Over time, we became more sophisticated and surgical in focusing on specific geographical locations and sectors of the economy. In developing the system of protection, the initial option was imprecise. Some pundits alleged that the administration was only elevating the threat level for political purposes, but I can a.s.sure you that in each case we believed that the threat was real and imminent and that we had no other reasonable option.

While we raised the threat level on four occasions during my tenure, one period stands out in my mind: the spring and summer of 2004. There were several streams of concern. First, we came into the possession of casing and surveillance reports focused on financial inst.i.tutions in New York, New Jersey, and Washington. What was noteworthy about the reports was their specificity and attention to detail regarding the buildings themselves, perceived structural deficiencies, the location of security, and the types of alarms in specific locations within the buildings. The reports were written as though produced by an engineering consulting firm and were of a quality consistent with what a sophisticated intelligence service might produce. Only one dot to connect, perhaps, but there were more.

The strategic context for concern was compelling. We were approaching national political conventions and an election. Al-Qa'ida had paid attention to the fact that the March 11 attack in Madrid had brought down the Aznar government in Spain. We believed that Bin Ladin himself had a.s.sessed that a logical time to attack the United States was just before the U.S. election, when he perceived the uncertainty created by a potential transition of government would make a response more difficult.

There was the fear that the arrests of operatives in Canada, Pakistan, and New York suspected of planning attacks in London might force al-Qa'ida to accelerate the timing of attacks inside the United States. Because of military operations conducted by Pakistan in the southern tribal areas of Waziristan, al-Qa'ida was under enormous pressure, stimulating the need for a high-stakes showdown with the United States. The plotting against Musharraf's life continued.

The intelligence that we received was more frightening. By July 2004 we believed that the major elements of the plot were in place and moving toward execution and that the plot had been sanctioned by the al-Qa'ida leadership. We believed that al-Qa'ida facilitators were already inside the United States, in an organized group-which to the best of my knowledge has never been found-and that they had selected non-Arab operatives to carry out the attacks.

A separate stream of reporting told us of al-Qa'ida plans to smuggle operatives through Mexico to conduct suicide operations inside the United States. This was linked directly back to direction being provided by al-Qa'ida's leaders. All of this was consistent with the intelligence dating back to 2001 of either the presence of, or attempts to infiltrate, operatives inside the United States.

There was strategic warning, further arrests, and disruption activities overseas and in the United States by CIA, our foreign partners, and the FBI. NSA was operating at a fever pitch attempting to determine linkages from dirty numbers overseas to numbers inside the United States. Detainees were questioned and financial data mined for operational activity, all in real time. We posited likely targets and methods of attack. It was a period of furious activity.

The attacks-based on very credible reporting-didn't happen. Why? Had the effectiveness of law enforcement and intelligence disrupted the planning? Quite possibly. Was it a conscious decision on the part of al-Qa'ida to delay for its own reasons, out of concern for its weaknesses and the rally-round-the-flag impact an attack would have in the United States? Equally plausible. It was yet another period of high threat that had not come to much, other than exhaustion. I do not know why attacks didn't occur. But I do know one thing in my gut: al-Qa'ida is here and waiting.

The threat was not just within the United States. Often information I heard at the five o'clock meeting would cause me to schedule abrupt overseas trips to key Middle East capitals. At one such meeting, I learned of intelligence that al-Qa'ida operatives were planning to a.s.sa.s.sinate members of the Saudi royal family and overthrow the Saudi government. I quickly scheduled a meeting with the Crown Prince.

ThenCrown Prince Abdullah is an incredibly impressive man, a billionaire like many Saudi princes, yet one who has never allowed himself to forget his roots. Alone among the top royals, he'll go off and live in the desert for weeks on end to reconnect with the Saud family's past. As cooperative as he could be in our pursuit of intelligence on terrorists, from our perspective, Saudi cooperation against al-Qa'ida could be slow and frustrating.

The Saudis were equally frustrated with us for not sharing enough information, but the speed with which we needed Saudi action came only after the kingdom itself was attacked in May of 2003. Thirty-five people, including ten Americans and seven Saudis, died, and more than two hundred were injured in the al-Qa'ida attack on a Western housing compound in Riyadh. That brought the message home to the royal family in a way nothing else had.

When I first heard about the Riyadh attacks, I knew I had to go see the Crown Prince, to offer condolences and to make a point while the wound was still fresh. I cleared the trip with the president and the national security advisor and gave them a rough idea of what I was going to say. But I wrote out my own talking points for use with the Crown Prince, and I didn't clear them with anyone. There was no reason to do so. I knew what had to be said. I doubt if I've ever had a more direct conversation with anyone in my life.

First, I started with an intelligence briefing on what had just occurred:

* The debate within al-Qa'ida over conducting attacks in Saudi Arabia dates back to the fall of 2002. It was never about whether to strike, but about when and how.

* The loss of sanctuary in Afghanistan, the settled areas of Pakistan, and northeastern Iraq raised an important question: Could the group afford to lose its position in the kingdom and, with it, its chief source of funds?

* Bin Ladin, who prior to 9/11 had imposed a ban on attacks in Saudi Arabia, made his position clear when he urged a key Saudi-based operative, Abu Hazim al-Sha'ir, to move forward with the attacks at any price.