by Mark Riebling

A Year of Transition

Army Lt. Gen. Donald V. Bennett, named DIA Director on 9 September 1969, was no spy. He was a warrior. In World War II, had commanded the 62nd Armored Field Battalion in the landing at Normandy. After the war, he had been an artillery commander with I Corps in Korea, then Superintendent at West Point. When summoned to DIA, at the age of 54, was commanding VII Corps in Germany. His lack of intelligence background not considered a liability, but an asset. To get a top-quality officer who'd be respected by the Joint Chiefs, Froehlke had urged Laird, they must recruit not from the "bespectacled weenies" in MI, but from the ranks of the "real men," who had led troops into battle. Laird had at last agreed, and Bennett dutifully accepted -- along with his third star -- the most daunting mission of his career: reorganizing DIA into "a credible producer of national intelligence."

There followed what Bennett humbly called "a year of transition." He had been "a user of intelligence for over 20 years," as he liked to say, but never "a supplier." Though DIA, under Bennett, would branch out into the foreign-policy and political jungles of Europe and Asia, Bennett himself was bewildered by his new home. He sometimes showed up late for meetings because he got lost just walking around the Pentagon. Adapting to headquarters from the field, while trying to "learn rapidly the responsibilities and relationships of agencies," presented what he candidly called "a personal problem." He found himself saying often, as he did when asked by Congress about DIA's Computer Assisted Subsystem: "Sir, I do not recognize that." 

Every Friday morning, from 7:30 to 9:00, Bennett met privately with Laird. "So, General, how's it going?" Laird would at ask over coffee. The meetings had something of the character of a graduate student briefing his advisor. But the two were learning together, and Laird was supportive always. He liked to pat Bennett on the back as went out the door. "The Chiefs giving you any crap?" he'd ask. "Not any more than usual, Sir," Bennett would ritually reply.

By 1970 Bennett was getting a feel for the "operational environment." He would pop into some of DIA's offices unexpectedly, if sometimes by accident, and poke around. He would talk up the working-level guys, ask them about their work. When the ANSRS network came online, he sat down bravely in front of terminal and asked for help. Then he marched over to Capitol Hill and told Congress why the system was worth fourteen million dollars:

"Suppose you have an aerial photograph of the airfield at Hanoi, and all of a sudden one of your photographs picks up what appears to be three or four bunkers on the side. You do not know whether they are underground revetments, ammunition storage areas, or what. You would go into your data bank" -- and so on, in detail. The practical, tactical emphasis gave Bennett's testimony a clarity, and force, that Carroll's had been felt to lack. And ANSRS, the Congressmen had to admit, sounded "pretty neat.''

But Bennett himself was by no means beguiled by technology. He believed more in human beings, and made it his business to bring better ones to DIA. The key, he decided, was not to raid the services, but rather to bring in -- and promote -- more civilians. The move surpassed some, given Bennett's military background, but was entirely in keeping with the essential "we can all learn from each other" humility of his character. Bennett had vowed to make DIA a "credible producer of national intelligence," but Lyman Kirkpatrick, who sat on the board of the Defense Intelligence School (DIS), pointed out by 1968 that "DIA is not equipped for the preparation of national intelligence, because by its military nature, DIA could never attract the type of civilian intelligence officers that have manned CIA... [who could] go deeply into the political background of various nations, complicated international arrangements, and the psychology of world leaders, more of whom are civilian than military." The Joint Chiefs themselves had recommended to McNamara, way back in July 1996, the "substitution of civilians in certain DIA posts occupied by military personnel. But McNamara had not felt he had anything to learn from the chiefs, and after glancing for a few seconds at DIA's employment rolls, said, "Thirty-nine point-six-five percent civilian. Let us call it forty. That is sufficient." Intangible, and missed in the calculation, was that many non-military staffers were but "half-breed civilians" -- middle-grade service retirees who blocked the careers of true civilians, younger and more energetic and with degrees from good schools. Talking to some of these young civilians, Bennett grasped that most were not being properly motivated. "All the command jobs were held by military officers," he recalled. "Therefore, if you wanted to rise in the organization, you never could get to the top spot."

General Bennett changed that, by what amounted to affirmative action. "One of my goals," he told Congress, "is to have civilian chiefs for four of DIA's eight directorates." When Bennett cut down on the number of people reporting to him, to reduce the intellectual clutter, he did so by cutting military men out of the loop, so that half those who got daily "face time" with him were civilians. When civil-service hires had sufficient time-in-grade, moreover, they were automatically promoted, while military officers had to be considered by a special board. By 1972, word had got around government that, at least for civilians, "the best-paying and fastest-promoting agency in intelligence is the DIA." 

But Bennett, to his credit, saw that morale would need more than merely material remedies. Good workers had to be inspired by a belief that their work itself mattered. General Carroll, too, had seen this, and in 1969 had asked DIS to update its wargames and training scenarios, and at least "change some of the names overtaken by world events, to provide an "aura of currency and insure student motivation." But the problem ran much deeper. Carroll had lamented how

to the average person who receives his impressions from radio, television, newspaper headlines, and spy novels, the intelligence business is a bizarre and exciting adventure. We run into this when we attempt to recruit college graduates, and we have to be careful to clarify this mis-impression. Otherwise, they become disillusioned, and leave us, when they find out that the intelligence business is really the accumulation, sifting, sorting, and analyzing of literally millions of bits of information.

Lucky even were those who got to do the sorting-and-sifting, as was discovered by a West Point graduate whose real name, incredibly enough, was James Bond. Assigned to DIA's "Pentagon Liaison Office," Captain Bond never worked with intelligence at all.

Yet DIA, despite all denials, did do espionage. The agency did deploy what were popularly termed "secret agents," and some of their missions were bizarre and exciting enough. These operations, however, were both bureaucratically and semantically cloaked. Since the very first briefing to Congress, in 1962, the line had been: "We don't allocate any DIA funds for what would be called clandestine operations." In the narrow technical sense, this was true. The funds were drawn from service accounts, and from a "black" portion of the overall Pentagon budget (for which no public Congressional account was given). The funded programs were not officially called "clandestine" operations," but were referred to rather as "special collection by worldwide tasking from DIA," or "classified expenses for extraordinary military activities." These activities were, until Bennett came along, run "back-channel," by DIA's Counterintelligence and Security Directorate, under Pete Dirzes and Tom Fox. Compartmentation was so tight that when the chiefs of other DIA directorates testified to Congress, under oath, that DIA was not engaged in covert ops, they were honestly ignorant. Fox never testified, but did say later: "DIA was involved in black operations since its inception." Operation COLDFEET, the 1962 mission to Soviet ice-station NP-8, had been but the first of these secret endeavors. But not until 1976 would the agency publicly admit to "clandestine espionage, done with maximum security and concealment" -- a statement glossed with the confession, by one of the agency's former directors: "DIA has done some unwise things, I think."

Unwise, in the Pentagon, was any operation which did not work. The "take" from DIA's clandestine human programs, after eight years, was considered meager indeed. Though Dulles had at first feared rivalry "in the operational field," the perceived ineptness of DIA's continuing "experiments" caused CIA officers to eventually relax. They laughed openly, for instance, when they heard that

one DIA operation involved throwing masses of used whiskey bottles left over from officers' messes into the Danube. Anti-Communist messages would be inserted in the bottles, and when, carried on the Danube's broad bosom, they reached the Communist countries, the inhabitants, hopefully, would fish them out of the river, read the incendiary messages, and revolt against their Communist masters. This plan was finally abandoned after much debate -- no doubt fortunately, since it is hard to imagine anything more likely to instill anti-American sentiments than the disappointment of finding thousands of whiskey bottles filled only with inept propaganda.

Carroll had tried to fix things, late in the game. In 1968 he shifted "human field-operations" from Col. Fox's office to a new "command and control center" at Fort Belvoir, Virginia. This new Human Resources Division, operated by DIA's Collection Directorate, was commanded Army Col. Robert C. Roth. But the lack of any detectable improvement led then-SecDef Clifford to conclude, fatalistically, that "penetration is a practical impossibility today through use of the human agent."

General Bennett disagreed. Yet that was not why he wanted to involve DIA more deeply, and officially, in human espionage. Bennett wanted more "black ops," he told Laird, "even if they didn't work." When Laird with some shock asked him why, Bennett said simply: "We're a supposed to be an intelligence outfit. If we don't spy, our own people won't think we mean business. Can you imagine being a soldier in Army and being told you'll never fight? How seriously would you take your work? That's the kind of situation we're in."

Bennett searched, as he so often did, for a civilian solution. In 1970 he proposed expanding DIA HUMINT under the supervision of the National Foreign Intelligence Board, chaired by DCI Helms. The idea was bitterly resisted by CIA civilian case officers, who worried about losing turf to the military. CIA Counterintelligence Chief James Angelton worried, too, about widening the number of persons who might "need to know" about secret agents-in-place. But Helms set up an ad-hoc task force to study Bennett's idea. And while it was being studied, DIA finally proved it was ready to become a "made" a member of the black-ops Mafia. It helped carry out an assassination.

Death in Santiago

At two o'clock in the morning of 22 October 1970, in a quiet, dimly let industrial section of Santiago, Chile, DIA attaché Paul Wimert delivered three submachine guns, 500 bullets, and six gas grenades to an officer of the Chilean Army. The thing that happened, eight hours later, was but the last play in a secret game Bennett and Helms had begun playing thirty-five days before.

On 15 September, President Nixon summoned Helms to the White House and told him that Dr. Salvador Allende, a Socialist who had recently been elected President of Chile, was "not acceptable to the United States." Helms duly began to plan a coup, but CIA lacked contacts in the Chilean military to carry it out. So Helms turned to DIA.

"No problem," Bennett told him. By 28 September a priority collection requirement had been sent to Col. Roth, at Fort Belvoir, "to identify Chilean personalities who might be helpful in preventing the election of Allende as President of Chile." Though Allende had already been elected, Roth understood what was meant -- and why the request was so misleadingly phrased. He cabled Col. Wimert, at the Defense Attaché Office in Santiago, and asked for a list of potential conspirators. A list was provided. By 5 October Wimert himself was authorized -- in messages bearing Bennett's signature -- to "contact and advise the principal military figures who might deny the presidency to Allende." Over the next two weeks, Wimert met twenty-one times with associates of retired Chilean general Robert Viaux, who was considered "a general without an army." The purpose of these contacts -- which Bennett later called "routine DIA activities" -- was to ensure the death or disappearance of Allende's key ally, General Rene Schneider, chief of the Chilean armed forces.

"To eliminate General Schneider," Wimert cabled Bennett on 8 October, "to either replace him, send out of the country," or "to kidnap him," members of the Chilean War Academy had asked for "light weapons." They also wanted "several hundred paralyzing gas grenades to launch a coup." But a shipment so large would be hard hide, no matter many courier pouches were used. So ten days later came the scaled-down request for "8 to 10 tear gas grenades, three .45 caliber machine guns, and 500 rounds of ammunition." The conspirators had their own weapons, of course, but needed guns which could be "identified by serial numbers as having been issued" to them."

The guns, grenades, and ammo arrived from Washington, in a diplomatic pouch bearing the seal of the United States Department of State, on the morning of 19 October. Later, that same evening, the conspirators toasted General Schneider at special banquet in his honor, which was supposed to conclude with his kidnapping. The plan was called off because Schneider's police guard was too heavy. But three days later, the plotters struck in force.

Shortly after 8 a.m. on 22 October, while he was on his way to work, General Schneider's car was intercepted and forced off the road. He drew his handgun, but was dead before he could defend himself. Chilean police later found an empty machine-gun at the scene, but were unable to trace it.

Human Resources

The assassination of General Schneider did not cause the toppling of Allende. If anything, CIA concluded, the killing rallied the Army "firmly behind the flag of constitutionalism." Allende would remain in power until 1973, when the Chilean military decided he was not a socialist, but a communist, and killed him in a coup. But DIA had come through in Santiago, showed it could play the game. Within hours of Schneider's death, Helms pronounced approvingly: "maximum effort has been achieved."

If Bennett's conscience bothered him, he never let it show. He was, after all, a warrior. He could accept that people died in wars, and in secret wars by secret means. What he would find harder to accept was that these means, in democracies, did not always remain secret. When CIA's William Colby later told Congress what happened in Santiago, and turned over cables revealing DIA's role, Bennett denied ever seeing or signing off on the messages which bore his name. Col. Roth, and CIA covert operations chief Thomas Karamessines, both testified that Bennett had been fully in the loop. But Bennett stubbornly insisted that he lacked authority "for covert operations of that magnitude."

His covert-operations authority increased by several magnitudes, in any case, shortly after the killing of Schneider. Though there is no explicit proof this was a "reward" for DIA's performance in Chile, the sequence of events is suggestive. In June 1971, Helms' study panel approved a one-year trial of Bennett's covert operations plan. DIA's Human Resources Division -- watched by a Bennett-proposed interagency Human Resources Committee -- could now carry out missions to give "local area commanders a trained base of agents in time of war."

Bennett got busy in a hurry. Within days of the study panel's decision, a four-man DIA team covertly crossed from Iran into Afghanistan -- to "verify mineral reserves" in the Hindu Kush, and to photograph shipments of Soviet materiel stored in strategic mountain passes. Before 1971 was over, another unit began trying to entrap KGB officers in Bangkok, and sent agents into China through Hong Kong. By June 1974, after two annual renewals, the results were judged successful enough for Bennett's trial plan to be accepted on a permanent basis.

So the Bennett years were good ones for DIA. He was bold, he was prudent, but perhaps above all he was just plain lucky. Even as he was enlarging DIA's human-collection role, one of the crown jewels of American espionage fell also into his lap.

Senior Crown

In late 1969, the National Reconnaissance Office turned over control of all CIA spyplanes to the Pentagon. Though DIA had directed U-2 flights during the Cuban Missile Crisis, supervision had soon after reverted to CIA, and for seven years the military had made do with EC-121s and other converted conventional craft. But now they had the "hot rods," the rocket planes. They had not only the old U-2s, but the new-model SR-71, the fastest plane in the world.

They called her Blackbird. She could fly three times the speed of sound, at 80,000 feet, on the inner edge of outer-space. She would be attacked more than a thousand times, after her first operational sortie on 21 March 1968, but she would never be shot down. One of her pilots, Air Force Col. Richard Graham, would fondly recall "MiGs falling out of the sky from attempted intercepts as they ran out of airspeed and ideas."

Imaging sensors were packed in a removable nose. For panoramic photography, a high-resolution Optical Bar Camera scanned from left to right, continuously over diamond-shaped area, up to 100,000 square miles in one hour. Electro-magnetic sensors were also used, to ferret radar signals, though for this the plane's high speed was a drawback. The SR-71 flew so fast, the enemy's electronic environment was stimulated only briefly before she was out of tracking range. "Our missions were not generally ELINT productive unless 'they' were looking for us with electronic signals," Graham recalled. "It often made me wonder if there wasn't a more covert classified program in existence, that somehow gave advance notice of the SR-71's arrival, just so 'they' would have their electronic defenses up and running for us to document."

Within DIA, Blackbird missions carried the code-name SENIOR CROWN. The operations were paid for by the Air Force, but DIA decided when and where they were flown. They were supposed to be flown, for instance, for SIOP nuclear bomb-damage assessment, to determine whether follow-on strikes were warranted. Since nuclear war never occurred, of course, most missions were for strategic I&W, Indications and Warning. DIA would send the Blackbird over China and Russia on a regular basis, to monitor troop movements, aircraft deployments, and other standard indicators of imminent action. The plane also played a tactical recon role, notably in Vietnam, where she sought acquisition-radar signals from SA-10 anti-air emplacements. The NVA's Soviet advisers didn't turn the SA-10 radar on too often, knowing the signals would be recorded. But when they were switched on, the SR-71 would lock onto and record them, and the data would be sent back to the Pentagon, where DIA and Air Force analysts would work on jamming and other countermeasures. For all these purposes and others, SR-71 missions were flown from only two secret bases: from Kadena AFB, in Okinawa, Japan, under the code-name GIANT SCALE, and from Mildenhall AFB, England, under the code-name GIANT REACH. Each flight was monitored start-to-finish by DIA watch officers in the Joint Reconnaissance Center.

Bennett handled these flights wisely. As his control over airborne recon increased, he decreased the number of flights, cutting $80 million from Peacetime Airborne Reconnaissance Program (PARPRO) over two years. He also raised the bar for risk assessment, and made sure that fighters were always on "strip alert" for backup. Nor did he "fall in love" with the Blackbird. The U-2 production line was reopened, partly because Bennett believed the old "grey ladies," despite their "high attrition" rate, could still play an important role.

He was right. On 26 August 1970, a U-2 overflew the Bay of Cienfuegos, a deepwater harbor in Cuba, and came back with photographs that caused tremors at the highest levels of government. The imagery showed military barracks that had not been there during the last overflight, eleven days before. In the interim, Bennett had received ONI reports of a Soviet nuclear-attack submarine flotilla heading for Cuba. Were the Soviets trying again to base nuclear weapons ninety miles from the United States? Did a second Cuban Missile Crisis loom?

National Security Adviser Henry Kissinger thought so. When he was shown the U-2 photographs, he fixed immediately on "recreation areas" adjoining the new barracks. DIA photo-analyst John Hughes had determined the areas contained "soccer fields." Kissinger tapped excitedly on the black-and-white prints. "Cubans do not play soccer," he said. "Russians do!"

DIA's Latin American Division sent up a paper saying that Cubans did, in fact, play soccer. There were several-dozen soccer fields throughout the island, a check of DIA's Imagery Library revealed. But Kissinger could not be swayed, and leaked his own analysis of the U-2 photos to the press. The intelligence community itself was meanwhile divided on Cienfuegos. Bennett's analysts argued it would be at most a minor rest-and-recreation port for Soviet subs, because the barges and other infrastructure for long-term basing were "just not there." The Navy sided with Kissinger, and CIA tried as usual to stake out some middle-position. When Bennett briefed Congress about the situation, on 9 October, Senator Frank Church did not think DIA's intelligence could "sustain a reasonable conclusion one way or the other."

A crisis was averted when Kissinger extracted a Soviet promise not to base nuclear subs in Cuba. Nixon declared "diplomatic victory," but Kissinger's unilateral approach to the intelligence -- and his decision to confront the Soviets before all facts were in -- had raised in the spy world a good many eyebrows. He seemed to see himself as a sort of intelligence czar. He demanded to know the sources behind DIA reports, and when he disagreed with the reports they came back to Bennett with "piece of crap" scrawled across the first page. Above all, Kissinger wanted intelligence put at the service of policy, rather than policy shaped by intelligence. DIA's Vernon Walters would recall how, in one meeting, when Walters recommended a certain option based on the available data, Kissinger had wagged his finger.

"You in intelligence," he said, "don't have a vote."

Walters in Paris

Walters knew well how Kissinger worked, for in August 1969 he had become Kissinger's secret partner in a private, out-of-channel diplomatic-intelligence operation. This privatization of covert policy was later resented by the official intelligence community, especially CIA, as having a "serious and adverse impact." Helms' people, it seemed, had been working strenuously in one direction, while Kissinger was clandestinely undercutting them in another. But DIA, during Bennett's directorship, would not have as much reason to complain. If CIA was increasingly "out of the loop," DIA's emerging potential as a policy tool was tapped when Kissinger cut in Walters.

Walters had been DIA's senior attaché in Paris since the fall of '67, when his Vietnam fact-finding mission had been cut short by Carroll. As Walters had hoped, his time in Vietnam gave him an "in" with French military officers who had served there. But the French secret services were far less friendly, suspecting Walters had been sent to "stir up the Army on behalf of NATO and against General de Gaulle." They surveilled him, tried to entrap him sexually by "dangling" girls and boys. They tried to intimidate him, by showing him how deep were their dossiers: when he was eight years old, they reminded him, he had been arrested for "riding a bicycle without a license plate."

But Paris was Paris. Everyone wanted to be there, and sometimes it seemed to Walters everyone was. The Paris DAO was the largest that DIA ran, larger even than its stations in Moscow or Berlin. Under Walters were fully 30 staffers, with of course their wives, and most had yanked hard on Pentagon strings to get their posts. A great many senior officials also found excuse to visit from Washington, and since Walters' DAO budget covered only foreigners, he had to entertain at his own expense these Americans in Paris. Indeed, Walters was at times more preoccupied by problems of etiquette than counterespionage. When a chicken bone stuck in his throat during one elegant dinner with foreign dignitaries, he calmly excused himself from the room and went into a hall-closet, where he could stick his fingers in his mouth without seeming "a coarse American peasant."

Walters' most demanding and dangerous mission in Paris, until Kissinger came along, had been laid on by DIA during the French student riots. In May 1968, hundreds of thousands of left-wing activists had virtually taken over the city. Red flags, and the black flags of anarchy, were raised over occupied factories, office buildings, and some fine restaurants. De Gaulle at first refused to crack down, and there was speculation that the French military would get fed up and remove him. Walters could hear "the clanking of the tanks... in the middle of the night." The Pentagon wanted to know what would happen, so Walters sent cables to DIA's Indications and Warning Center, under the heading: "Impact of Current Events on the Armed Forces." At Walters' behest the Air Force attaché stood out on the window ledge, five stories up, scanning the Left Bank with binoculars, shouting out police and troop movements to an Army stenographer. Walters himself strode out into the chaos, in French clothes pinned with a Legion of Honor boutonniere. On one occasion he was overwhelmed by tear gas, and when he stumbled back to his car some rioters threw a Molotov cocktail, which sailed just over the hood and exploded nearby.

Despite all the violence, Walters assured DIA headquarters that "this rioting was not going to overthrow General de Gaulle." The students, he reported, were only alienating most working-class Parisians. When they denatured the tree-lined streets with chain saws, to make barricades, old women screamed at them about how long the trees would take to grow back. Though they proclaimed a "dictatorship of the proletariat," Walters perceived that most of the rioters were "rich kids," acting from "chic leftism and guilt for their inherited wealth." He related that when one of these privileged revolutionaries had tried to shock his father, by bragging that he had "set fire to seven cars" that night, the father said, "Well, you've got me beaten. I only set fire to one." The boy, caught off-guard, asked, "Which one did you set fire to, Dad?" Whereupon the father said solemnly, "Son, I set fire to your sports convertible." Walters thought that summed up the general attitude toward the rioters, and his view was borne out on 31 May, when the students were routed by a million ordinary folk singing the Marseilleaise.

The next seven months were quiet, filled mostly with social routine, but on 20 February 1969 Walters was called by urgent cable to the White House. He had known Nixon since 1958, had served as interpreter during the-then Vice President's swing through South America. They been trapped together, for a time, by angry "anti-imperialist" crowds in Caracas, and Walters had been impressed by his companion's calm. Nixon, for his part, had been impressed by Walters' capacity for writing "memcons," memoranda of conversations, without taking notes. Now, in the Oval Office, Nixon asked whether Walters would be interested in recording all presidential conversations. Walters politely declined, and Nixon not long afterward installed the tape recorder that would undo him. In the meantime, the President gave the impression that he might want to call on Walters for some other special, but unspecified, jobs. Walters returned to France without any idea what Nixon had in mind. But he found out in August, when Henry Kissinger showed up unexpectedly in Paris, and asked to see him.

On 15 July, Kissinger said, the president had sent a letter to North Vietnamese leader Ho Chi Minh, proposing secret negotiations to end the war. Minh had agreed, and Kissinger was to meet North Vietnamese delegates in Paris. He wanted Walters to help him, both to translate and "to serve as a channel."

Walters had supported an earlier, open series of peace talks in Paris, which had failed. The discussions had barely got past "the shape and position of the table at the conference and who would sit where." But supporting Kissinger, as Walters agreed to do, would be a much different business. No one in the Paris DAO, except Walters' secretary, was to know. He would have to organize, from scratch, what he correctly sensed would be "one of the great secret missions of our time."

Walters focused first on "tradecraft." He asked for and received NSA special codes for cabling Kissinger at the white House, and each agreed to identify himself by the code-name ANDRE if he had to make contact by telephone. All that was easy enough; more difficult would be getting Kissinger into-and-out-of Paris, without the world knowing. Walters decided to use Air Force one: it flew abroad occasionally on "training missions," its passenger lists were never made public, and it was exempt from customs or immigration. Kissinger would fly into various airports around Europe -- always at night, with his coat collar up and a hat drawn down over his face -- and transfer to a DIA plane for the flight to Paris. Since Walters' official vehicles had diplomatic plates, he had to rent cars for getting Kissinger around the city, and after a while the cost started adding up. But when he broached the subject with his co-conspirator, he was airily cut off. "I am trying to restore peace to the world," Kissinger said, "and you are burdening me with all these administrative details."

To Walters the details were less burdensome than the talks themselves, which were often anticlimactic and grim. The North Vietnamese delegates lived in fashionable the part of Paris, but insisted the meetings be held in a working-class district, in a shabby flat at 73 Rue Jules Lagaisse. There, watched only by a framed portrait of Ho Chi Minh, the two sides drank tea and had at each other in French. Kissinger would come in dozens of times, the talks would drag on for years, and at first Walters worried the work would lead nowhere. Then, in spring 1970, the North Vietnamese began suddenly to soften up. It seemed they were "feeling the pressure of the war."

They were indeed -- but not so much in North Vietnam as in Cambodia, where they had long operated sanctuaries. For even as Kissinger, through one man on the DIA payroll, was tendering truce in Paris, he was hitting the enemy with other DIA assets in the Cambodian capital, Phnom Penh.

The Parrot's Beak

Prince Norodom Sihanouk, ruler of Cambodia, had tolerated North Vietnamese sanctuaries along his borders since the days of the French defeat at Dien Bien Phu. He called this "neutralism," but Kissinger called it cowardice. Sihanouk had been feuding with Washington for years, and by 1970 he suspected the CIA was plotting his removal. He was essentially correct; he just got the initials of the agency wrong.

DIA, as Bennett came in, was trying to get a handle on Cambodia. Nixon had ordered the secret bombing of Cambodian sanctuaries (Operation MENU), and the Joint Chiefs needed targets. DIA's Mapping and Charting Directorate had been not quite up to the task, which in all fairness had imposed on them literally overnight, and much of the targeting was done with old Michelin road maps. Bennett was determined to get better data, and the Joint Chiefs gave him the chance. They wanted to know, specifically, about the "level of support" flowing to the VC through Sihanoukville, a built-up port with good roads to communist bases. Bennett had three estimates: MACV's, which was highest; CIA's, which was lowest; and his own analysts' judgement, which was in the middle. None of the numbers were hard, so Sihanoukville was for a time let be. But 110,000 tons of bombs did fall on other parts of Cambodia, and as Bennett tailored reconnaissance the targeting got more precise.

The bombings did not cause Sihanouk to shut down the sanctuaries, as had been hoped. By January 1970, Kissinger was annoyed with the prince's Machiavellian "expedience." The message to the Pentagon, recalled Kissinger's assistant Roger Morris, was an ancient phrase, and unmistakable: "Who will rid us of this troublesome prince?"

DIA began to make what Morris called "shadowy contacts" with dissident Cambodian generals. As in Santiago, four months before, and as for two years more in Paris, special use was made of the Defense Attaché Office in Pnomh Penh. Bennett's attaches, it seemed, had become the "instrument of choice" for Kissinger and Nixon.

But Bennett himself worried whether the DAO in Cambodia would be up to the job. While the office was comparatively large, its product had been sometimes mediocre. Attaches would visit the border areas, cable daily updates, but Congressional neutrality restrictions prevented them from accompanying Cambodian units into combat areas. The Pentagon, for its part, prohibited its personnel from entering the government-gossip center of Pnomh Penh, an opium parlor run by one Madame Chantal. So information was hard to come by, and one DIA analyst claimed the Cambodia attaches "made little contribution to the base of knowledge about the military situation."

The senior Defense Attaché in Cambodia, Colonel William Pietsch, had a lot to prove. Journalist Joseph Kraft, who got to know "Colonel P" in Phnom Penh, found him "a creature of comedy and maybe pathos, chiefly distinguished by the wearing of a white bartender's jacket." Pietsch was often overhead propounding a "reverse domino theory," insisting the enemy was "on the run." If only the Cambodian government were brought into Uncle Sam's camp, it would be the beginning of the end for communism in Southeast Asia. "I'll bet my professional reputation," Pietsch boasted, "that we'll bring it off."

Though perhaps "not a very effective diplomat," as Kraft flatly wrote, Pietsch proved to be a pretty good spy. On his watch, DIA got discreetly in touch with Cambodian general Lon Nol, who despised Sihanouk's appeasement of the communists. A DIA biographical report noted that Nol's mind sometimes ""tended to take flight," but that "he had been a friend of the West and cooperative with US officials during the 1950s." To preserve secrecy, DIA usually communicated with Nol through a "cutout" agent, who worked in the Indonesian embassy in Phnom Penh. According to one former intelligence officer, Nol was assured of U.S. support in a coup; according to another, Nol kept DIA advised of his plans to stage one. In early March, at any rate, Bennett and some other Washington officials flew in to Pnomh Penh, ostensibly to discuss "Cambodian neutrality," but almost certainly to talk with Nol about his move against Sihanouk.

The neatness of what happened next implied a careful scripting. On 11 March -- while Sihanouk was conveniently out the country, considering a Kissinger "neutrality" plan conveyed through Walters in Paris -- 20,000 Cambodians vandalized the North Vietnamese Embassy in Pnomh Penh. The next day, Nol unilaterally announced that North Vietnamese ships could no longer dock at Sihanoukville. On 17 March, Nixon told Kissinger "Let's get a plan to aid the new government." The timing of that instruction suggested close coordination with Nol, who had not yet made public his plans. Nol ousted Sihanouk in a bloodless coup 24 hours later, and appealed immediately for American assistance.

Assistance was given. While Sihanouk went into exile in Peking, the U.S. 11th Armored Cavalry Regiment went into Cambodia. The operation destroyed communist sanctuaries in a region known as the Parrot's Beak, but caused violent protest at American colleges like Kent State.

When protest spread even to Paris, things came weirdly-full circle for Walters. Attaché work in France had helped attaché work in Cambodia, creating more attaché work in France. When students took over the American college on the Quai d'Orsay, they demanded the Embassy dispatch someone to explain why the United States had invaded Cambodia. Walters went over in civilian clothes, stood up to speak in an auditorium, and was shouted down. One young American presented him with a parcel wrapped in newsprint. "It fell open," Walters remembered, "revealing a fetus."

Blood Bath

While the Cambodian episode stoked political dissent, by August 1970 the war itself began to "simmer down." Pietsch's reverse-domino theory, it seemed, might actually hold. After Nol's coup, Walters sensed that North Vietnamese diplomats in Paris "were truly shaken," revealing perhaps just how important Cambodia had been to their war plans. That importance was revealed also by Viet Cong documents captured in the Parrot's Beak, which showed that DIA's estimates of the Cambodian scene had been dangerously off.

"After the fall of Prince Sihanouk," Bennett said, "we were able to look into the situation firsthand, and found all of the supplies in the IV Corps area [in the Mekong Delta of South Vietnam] came through Sihanoukville. Our figures showed that we missed it [the correct assessment] by almost a half."

Bennett's analysts miscalculated also the effects of the U.S. policy on Cambodia's own communist rebels, the Khmer Rouge. Sihanouk had aligned with them after his ouster, and the North Vietnamese had begun training and supplying them in force. When Nixon withdrew U.S. troops from Cambodia in July, the Khmer Rouge filled the vacuum and began implementing their infamous "blood bath" theory of social progress.

The Khmer had been around since 1951, and ignored for just as long. At the time of Nol's coup, both DIA and CIA had dismissed them as a rag-tag band of "five thousand to ten thousand." But CIA analyst Sam Adams, who liked to pick fights over numbers, began to question this figure. He dipped into the documents captured in Cambodia, which were overwhelming CIA's filing system just as VC papers did DIA's three years before. He went through cartloads of brown, cardboard boxes in CIA's first-floor archives -- studied chains of command, radio nets, hospitals, training camps -- and concluded, from the scope of this infrastructure, that the Khmer Rouge were "very big." Whence, he wondered, had the small numbers come?

At five o'clock in the afternoon of 28 May 1971, Adams called DIA and asked for the "Cambodia shop." A Sergeant Reisman came on the line. The numbers, he said, had come from Lon Nol's government a year before. Reisman had left the figures alone, because they seemed "reasonable," and DIA had never done an in-depth study. "There's no one at DIA to do a study," Reisman groused. "They hit us with too much shit."

Froehlke's Folly

General Bennett was hit with his own share of it. In his annual testimony to the Appropriations Committee, he was criticized for the "failure to understand the importance of Sihanoukville," which Rep. John J. Rhodes (R-Ariz.) called "one of the greatest failures of the intelligence community I know anything about." Rhodes was either exaggerating or ignorant -- Cambodia was nothing like Pueblo or Tet -- but Bennett acknowledged that "it was a failure in part." When he noted that he had, at least, forwarded MACV's higher (and essentially correct) Sihanoukville estimate to the Joint Chiefs, Congressman Whitten said, unfairly: "It sounds like you just pass on what the various agencies say." But Carroll had been treated far less courteously, and Bennett earned some good will by cutting $2,347,000 from DIA's budget his first year.

Arthur Froehlke did not fare as well. Laird's Assistant Secretary for Administration was supposed to be managing DODINT "resources," but seemed only to want to spend what Bennett would spare. He advanced, for the first time, an argument that would be often heard in later years: viz., that reductions in defense spending should entail bigger intelligence budgets. "As our forces go down," Froehlke said, "the risks necessarily go up.... I believe it follows that we had better know what is going on in the world, so that we can better utilize the lesser forces."

Whitten could smell a meal. As he so often did, he began with collegial, circling generalizations, wrapping Froehlke like a python, then bit where he wanted.

"I believe Mr. McNamara put the whole Defense Department on the basis of cost-effectiveness. I understand that is your job also, at least in intelligence. All that is to the good, but we do not know much about the cost-effectiveness part of it. You have meanwhile set out to get yourself a staff. This may sound funny, but what have you done?"

"I have put in a lot of hours."

"And?"

"We have improved the leadership and morale of DIA."

"How do you know?"

"Unfortunately I feel rather than see it. I can give you no examples."

"That makes it easy to say, when you don't have to prove it. It is so easy to sort of 'run out the clock' in this hearing, but we need some specific answers on these questions. What else?"

"I am working on a comprehensive inventory of DoD intelligence assets. No one has put all of these programs together and asked the simple question: How many collection platforms are focused against the same target? Today I am not in a position to tell you, for instance, how many platforms are targeted against the Soviet forces in Eastern Europe. I do not know."

"The Defense Department is supposed to have had somebody doing what you are doing from the beginning of time. Whose responsibility was this?''

"If you read the charter of DIA, you can read that DIA was supposed to perform this overall management. It did not do it."

"Why didn't you go after them and make them start discharging it?"

"In my opinion, we had serious problems in DIA. I hesitate to go into too much detail because -- "

"We have heard lots of things. It would be hard for you to shock us. Go right ahead."

"There are a lot of things that are, not critically wrong, but wrong with DIA. I don't think DIA would have been accepted as a resource manager."

"DIA, having been set up to pull these things together, having failed miserably -- why did you not abolish it and replace it?"

"I am strictly resources. I do not get into that part of it. But I -- I am going to dream a little bit with you, if I may. I can conceivably see my shop working itself out of a job. I can see -- two, three, four years from now -- being able to assign our responsibilities to DIA, saying, 'Yes, DIA is doing a good job; they have good people.'"

"Mr. Secretary" -- Whitten was at the jugular now -- "I declare you have not touched on a single one of the problems that have stuck up their ugly heads, but are in a dream world, trying to figure out on paper what looks and sounds good... [I understand] you are still having to take the time to get acquainted with the heads of departments. They have to report here, and report there, and little is accomplished. It generates a great deal of lost motion! These are busy folks. They don't have the time to sit around and shoot the bull while you throw questions at them. I suspect that they are in better shape to throw questions at you."

"I think," said Froehlke feebly, "that you are right."

This was in March 1970. Four months later, after furious work, Froehlke thought he could finally answer the big question, the one as old as the Pentagon: How much did defense intelligence cost? He had created what he called the Consolidated Intelligence Resource Information System (CIRIS). This was to be a sheaf of papers, with two columns of text. On the left would be listed all the collection platforms; on the right, the target and cost of each. It would be just like an actuarial table -- and Froehlke, of course, had been in the insurance line. "When we get all through," he had promised Congress, "we will be able to, as it were, push the button and say, 'What have we got?'" But Froehlke pushed the button in July 1970, he had to admit, what came out was "a rather primitive effort in terms of what we hope to get." Listing all the platforms required "working from the ground up and drawing on all sources." And how could one know, in the end, whether all the sources were known? If that riddle could have been unraveled in four months, it would not have been now in its fourth decade. Froehlke began to understand why the traditional symbol of military intelligence was the sphinx.

His efforts, in any case, were overtaken by events. A year before, just when Laird hired Froehlke, President Nixon -- always fond of parallel approaches -- had formed a "Blue Ribbon Panel" to investigate defense intelligence. The panel was chaired by another insurance man, Gilbert Fitzhugh, CEO of Metropolitan Life, and its report (1 July 1970) was written by University of Chicago economist George Stigler, who later won a Nobel. Stigler did not even try to cost the "Gross Intelligence Product," but instead asked why the always-great supply of intelligence never seemed to meet demand. He drew up a flow-chart, and what he got was a maze. "The principal problems of the DIA," Stigler decided, "can be summarized as too many jobs and too many masters." The agency was doing almost everything, for almost everyone, and almost never well. There was not enough division of labor, because arbitrary regulations prevented the labor from dividing itself. "The obvious answer to this bureaucratic mess," Stigler concluded, "is that the DIA should be reorganized along realistic, functional lines. Given the proclivity of bureaucracies to perpetuate themselves, however, this is no easy task. It would require bold, decisive action.

Words to make a soldier jump. Bennett had been crouched and ready for a half a year: he had ordered a "comprehensive review of DIA's organization" in early 1970, because "shortly after taking command of DIA it became apparent to me that something had to be done." When Stigler's report landed on his desk, Bennett moved with characteristic swiftness; by the end of July, he had torn down a ten-year-old empire and built a new one. He sorted out DIA's people and grouped them "functionally," as Stigler had urged. In the Production Directorate, for instance, a new Soviet Warsaw Pact Division was subdivided into air, ground, naval, missile, and other specialized shops. The data-flow was horizontal within each division, and vertical from each directorate to the director. The deputies in charge of the directorates were given more autonomy to act "on matters that were of primary concern to them," and Bennett hoped this would eliminate the "red tape and unnecessary administrative controls that had tended to bog things down." Carroll had wanted to sign off on everything, like his old boss J. Edgar Hoover. But Hoover had never been a soldier, and Bennett would never a be a bureaucrat. He junked this "FBI-type system" and looked for ways to reduce his own "span of control."

He even lopped off a whole directorate. When Congressman Minshall had complained that "everybody is in the mapping racket [and] they are all carried away with it," Bennett had not disagreed. Stigler recommended consolidation in a new Defense Mapping Agency, and on 5 November 1971, Nixon asked Laird to create one. Ten days later, the JCS approved Bennett's plan to transfer his mapping shop to DMA, which DIA would support with aerial collection.

Cutting into collection itself, however, caused sparks to fly from the ax-head. Stigler had found DODINT collection "excessive," and Bennett had thought the numbers of collectors to be so. Besides his Assistant Director for Collection, ten other officers reported to him in ELINT, TELINT, HUMINT, etc. He decided to put all these functions under just one man, veteran photo-analyst John Hughes, who would be DIACO for the next nine years. But Bennett still saw "areas of obscurity and differing interpretation in matters of who should do what," and Hughes' shop was still criticized for overcollection. The machines, the platforms, seemed to generate their own requirements: there would always be a tendency, as Froehlke perceived, to "develop options made available by rapidly expanding technology, simply because they are available." The machines were expensive, they were top-of-the-line, and their operators wanted to use them. Validated collection-requirements therefore tended, like cancer cells, "to acquire immortality."

Bennett sliced at the growth the only place he could, where it was human and soft. Carroll in his last months had asked Congress for $1 million for 41 more attaché personnel, but Bennett in his first months sacked 38 attaches for incompetence. By March 1970 the Defense Attaché System had lost fully 43% of its people, half its air-fleet, and offices in Cameroon and Somalia. Dollars and bodies were shifted to higher priority environments, especially Russia. The senior attaché in Moscow had been a colonel; he was replaced with a general. Bennett acted moreover to "insure, 100 percent" that each newly appointed attaché would be promoted to the next higher grade," and hoped this more prestigious career path would attract higher quality nominees.

"And on the seventh day, he rested." That, after Bennett's reorganization, was the joke that went round Arlington Hall. But he rested only briefly. Improving human collection, he felt, would mean little unless he could improve also human analysis. Since its earliest years, DIA had been dogged by the reputation that is threat assessments were "not objective." Its estimates of enemy strength were deemed, by the rest of the intelligence community, inherently parochial and excessively hawkish. The standard accusation was that DIA's parent organization, the Pentagon, wanted to justify new weapons and bigger budgets, and that DIA estimates were slanted accordingly. "That was always a problem," acknowledged Col. Tom Fox, who was in DIA but outside the estimates process. Fox did not think the problem was any worse or better than at CIA or State, which had their own reputations for slanting things -- e.g., to support foreign aid or arms control. But DIA had undercut its credibility with a lack of commitment. Rival agencies put their brightest minds in analysis, Stigler pointed out, while DIA's best people usually went into technical collection. Lowly analysts like Sgt. Reisman, in the Cambodia shop, lacked time and support for "in-depth" projects." Final judgments were increasingly made in other agencies, and "in order to avoid giving offense," as one critic wrote, the work of DIA analysts had become "bland and unadventurous." Bennett met that last criticism, at least, dead-on -- choosing from one of those other agencies a new chief analyst who was offensive, colorful and brash.

The Brash Advocate

Army Brigadier General Daniel O Graham, West Point '46, stood out in the intelligence crowd because he did not "project prudence." He was seen as "an energetic enthusiast," a "vigorous, very bright officer" who believed "every moment was a battle for the world." He was a specialist in Soviet affairs, and in his own view on expert on them: "I think I can say without fear of contradiction I was the dean of military estimators." Before Bennett made him DIA Deputy Director for Estimates, in November 1970, Graham had worked twice in CIA's Office of National Estimates, where he was remembered as "the funny little military officer who hung a drawing of a bayonet over his desk with a caption describing it as 'The Weapon of the Future."

Graham wanted to shake off the shackles of DIA's past. Because of earlier errors, he believed, the Department of Defense had become overly "defensive" about its estimates. For ten years, inflated calculations of bomber-gaps and missile-gaps had "hung like albatrosses around the necks of military intelligence officers." As a result, too many analysts wanted to "run with the pack," and too few wanted to "cry wolf."

Graham persuaded Bennett to pen-off the estimators from other the "Dogs of War." Estimates had been previously done in the Production Directorate, which was also charged with "answering the daily intelligence mail." Estimators were too close to commanders, too close to the pressures and politics of "the moment." Estimates were essentially a "scholarly problem," and would be better attacked in environment that was more "ivory tower." The new directorate would be located not in the Pentagon, nor even in Arlington Hall, but in Rosslyn, Virginia. There, Graham brought together widely scattered library and reference facilities, and set up shop with about fifty analysts -- mostly from DIA, but with some from CIA. He created a Key Intelligence Questions Program, to sharpen requirements for Defense Intelligence Estimates (DIEs), and demanded the new scholarship be published with report covers that were either blue, for "info," or red, for action.

Some of the older analysts couldn't hack it. Graham gave everyone a chance -- but only one chance. He had seen already the need for a "new crop of analysts and estimators." He wanted colonels with graduate degrees, who knew about Vietnam, not "Iron Majors" from World War Two, who seemed to all be middle-brow masons. He was tired of malapropism -- tired of reading about "labor materials breakout [sic, breakdown]," "contractural [contractual] assistance," "textural [textual] material," "Grifiss [Griffith] Air Force Base," "menstruation" [measurement]." He was sick of reading sentences like:

The great disparity [of] collection systems versus intelligence processing... has now reached a platitude [sic] where the anticipated payoff of a high cost collection system is limited by the DIA's capability to exploit them [sic] fully.

To Graham this was the prose of "a substitute mathematics teacher, which was just what one analyst became after the "selecting out." Graham used words like nomini universali, and he wanted to be understood. He wanted to "get academia aboard." He brought in people like Bruce W. Watson Sr., a Navy commander with a doctorate in Russian area studies from Georgetown, and Lt. Col. William Murphy, a Harvard linguist who had fought in Vietnam.

Graham wanted minds strong enough to break what he called "the greatest barrier to sound analysis: conventional wisdom." He wanted estimates to be "thoroughly scrubbed internally," by an "adversary process," before they were defended in the national arena. He encouraged academic sabbaticals, circulation through other departments, so that estimators would not go "stale" from "bureaucratic routine." Above all, Graham sought "questing spirits," who would reject "the most dangerous belief: that if information on a particular item satisfies the official intelligence requirements for that item, then we know all we need to know about it. Truth must not be "bureaucratically constituted."

He ruffled feathers, but earned respect. CIA officers conceded that Graham "quickly established the DIA office as a serious rival to CIA's estimative function." Accusations of bias abated, especially after DIA's main analyst on the Soviet Backfire bomber -- an Air Force officer -- downgraded Air Force assessments of the bomber's capabilities. Asked about prospects for a South Vietnamese invasion of Laos, Graham predicted fierce resistance; after the invaders were severely mauled, even Congressman Whitten praised DIA for "a good judgment of... the actual situation." Asked about North Vietnamese intentions, in June 1971, CIA ultimately selected DIA's analysis over its own. "If you are going to do all this," an impressed Whitten asked Bennett that same month, "why do we need the Central Intelligence Agency?"

Though the question was perhaps only rhetorical, DIA's improvement did, at last, lead Laird to ask: "Why do I need Arthur Froehlke?" Reforms had come despite rather than because of Froehlke, who was "not impressed with the reorganization method," and had recommended against "a radically new organization." He was replaced (3 November 1971) by a new Assistant Secretary for Intelligence, Albert C. Hall, who would "generally supervise" defense agencies. In the end, Froehlke granted that it was really General Bennett, and his "good leadership," that improved DIA.

"Let me make a flat statement," Froehlke said in his final report to Congress, in June 1971. "DIA is not the same agency it was two years ago. Recently, officials who inspected DIA asked the Joint Staff to evaluate DIA's support. Responses ranged from 'excellent' to 'absolutely outstanding.'" Where Froehlke had once called DIA "timid and reactive," he now sensed it was "taking the initiative."

It was indeed. When CIA seemed uncooperative, during the planning of a POW rescue-raid, DIA Captain Robert Harris threatened a Deputy Director of Intelligence at Langley: "If I find out you've held back one dot of information that will help free those prisoners, I'll personally commandeer a cruiser, sail it up the Potomac and blast this fucking building off the map."

The Raid

It was perhaps only fitting, given the agency's aggressive new spirit, that DIA should play a leading role in the first modern unconventional military rescue operation, which aimed to liberate POWs from a camp only two dozen miles from downtown Hanoi, the most heavily-defended city in the history of warfare.

The Pentagon had tried before to rescue American POWs, but always failed. Ninety-one missions had been mounted between 1966 and 1970, but only one American had ever been rescued, and he had died from injuries before he reached freedom. Not a single American POW had been rescued since the Civil War.

When General Bennett came in, DIA had yet to even locate any POW camps outside Hanoi. Bennett made the task of discovering one among his ten top-priority missions, and overhead coverage was targeted on "every building and compound in North Vietnam that had a wall around it." Thousands of photographs were taken, but all DIA got was black-and-white glossies of pigsties and water-buffalo pens. When finally there was a breakthrough, it came not from the lenses of on-site machines, but from the mind of a man half way around the world.

Vernon Walters had been searching for clues to North Vietnamese prisons for two years in Paris. France had occupied Vietnam for almost a century, and had built most of the installations that were big and strong enough to hold large numbers of men against their will. Through his connections in the French military, Walters gained access to archives and found "hundreds of documents... that helped explain the Communist prison system." Although some in the Pentagon dismissed the data as "outdated," Walters' tips allowed Bennett to focus SR-71 coverage on just a few sites, and finally led to the location of a prison.

In May 1970 an SR-71 flew over a site called Sontay, and returned with photographs of what seemed "coded messages" from American prisoners. Spelled out in laundry, rocks and dirt were the letters "SAR," the designator for "search and rescue mission," and "K," which stood for "come get us." Follow-up reconnaissance, in Bennett's hands by 2 June, confirmed the presence of "someone" in both camps. Army Brig. Gen. Donald Blackburn, JCS Special Assistant for Counterinsurgency and Special Activities, asked DIA to propose a rescue mission.

In a secure room at Arlington Hall, a fifteen-man team began to plan an operation that was at first code-named IVORY COAST, then POLAR CIRCLE, and finally KINGPIN. The plan, completed by 10 July, called for commandos to helicopter into the camp from Udorn, Thailand. They would be guided in by special C-130 COMBAT TALON aircraft, equipped with new forward-looking infrared radar (FLIR). The approach would have to be timed precisely to coincide with the swinging of enemy radar dishes, which must be passed while pointing in the other direction, or else the North Vietnamese warning system would "go hot." The raid would have to take place also under cover of night, and the moon would have to be "in a certain place," behind and to the side of the force, or it would "light them up." The lead helicopter would destroy Sontay's guard towers; a second chopper would crash into the compound and unload raiders, who would kill any North Vietnamese guards, liberate the POWs, and link up with two more helicopters outside the walls. Covered by A-l Skyraiders, the whole team would take off and to return to Thailand.

The Joint Chiefs said "go." Army Special Forces Col. Arthur "Bull" Simons was named operational commander, and though he had recently undergone heart surgery, he assured Air Force Brig. Gen. Leroy Manor, who was in overall charge, that he was back up to 250 push-ups a day. At Fort Bragg, North Carolina, on 13 July, Simons requested volunteers for "a classified mission involving considerable travel and risk," and about 500 men stepped forward. Fifty-eight were chosen for training at Eglin Air Force Base, Florida, where from September to November Simons made 368 practice assaults, supported by over 1000 hours of flying, and by at least as many man-hours of mission intelligence from DIA.

Coordinating intelligence for such a big and important job would probably have been provided in earlier years by CIA, but the Joint Chiefs were by now more comfortable using "their own" agency. Collecting and analyzing all-source data about the target was the exclusive duty of DIA, which was to draw on CIA and NSA for any needed support. Over the summer a "nap of the earth" flight-path was mapped, and NSA's Milt Zaslov found 5-minute gaps when the enemy radar dishes turned. DIA analysts noticed what seemed to be a school along the attack route, and warned Simons not to mistake it for Sontay. General Graham orchestrated coverage with high-altitude satellites and SR-71s, as well as low-level BUFFALO HUNTER drones, which gave a detailed view of the camp. John Hughes calculated the height of trees and walls, identified five main buildings and three guard towers, and pointed out where telephone lines might interfere with helicopter blades.

From Hughes' analysis, two detailed models of the camp were built. One, code-named BARBARA, was scaled to a table-top, and for $60,000 was fitted with adjustable lighting and eye-scopes, so Simons' people could see how the target appeared at different angles and times. The other mockup was full-sized, built in sections like theater scenery, and was used for live-fire rehearsals. DIA watch officers at DEFSMAC alerted Eglin whenever a Soviet Cosmos 355 satellite was about to pass over, so that the big model could be taken down and stored in a hangar. Even the post-holes were filled in, so that curious Soviet analysts could not "lay out" the objective.

There was only one problem with all this. While the fake camp bustled with activity, the real one seemed suddenly to have been abandoned. Reconnaissance after July showed few signs of life, and in August DIA reported the compound was probably "empty." It still seemed empty when an SR-71 flew over in October, but a flight the next month showed new crops, cut grass, and laundry hung to dry. Other intelligence indicated that while Sontay was back in use, it was no longer a POW site. By one account, this information came from a DIA agent in Hanoi, a "usually reliable source with direct access," who said the prisoners had been taken to a new camp at Donghoi. In mid-November, General Bennett told the Chairman of the Joint Chiefs, Admiral Thomas Moorer, that DIA could not confirm the presence of any Americans at Sontay.

Moorer was in no mood. By now Simons' commandos were already in Thailand, coiled for action. In a midnight conference at Arlington Hall, Moorer's deputy, General Blackburn challenged DIA's "ambiguous" assessment. "Look, you clowns, don't waffle it... All I want out of you guys is an answer -- at five o'clock tomorrow morning: Are they there, or aren't they? That's all." In essence the analysts were asked to "prove a negative," to prove the prisoner-transfer report was "not wrong," and of course they could not. Moorer decided to proceed, and on 18 November briefed the White House. Both Kissinger and Nixon had known of the plan since July, and final clearance was now given. A "red rocket" (flash execute) order reached General Manor the next day in his command headquarters at Monkey Mountain, just north of Danang.

Three hours before dawn, on 20 November, the raiding party began its run from Thailand to Sontay. The flight schedule was kept virtually to the second, and the force passed undetected through five-minute radar-dish gap. When they were 3.5 miles from the target, a C-130 dropped flares. The lead helicopter went in low between two wooden watchtowers and the door gunners opened up with 7.62 mm cannon, collapsing the structures instantly. The second helicopter crossed into compound over the west wall, with her door, window, and ramp gunners all firing, and executed a "controlled crash." The crash turned out to be not so controlled -- trees in the landing zone were higher than when last photographed, and touch-down through the branches was so jostling the door-gunner was thrown clear -- but the team was able to move out in squads. One squad blew a hole in the prison wall with a twenty-pound satchel charge. Another demolished the power generator with an anti-tank weapon. A third group advanced through a deep drainage ditch, cut through concertina wire with bolt cutters, and assaulted the main cell-block. They killed two North Vietnamese soldiers in the doorway, ten more inside, and began searching the cells with electric miner's head-lamps.

The rest of the force laid down covering fire. Five NVA rushed from one garrison, and were machine-gunned; four more were cut down running between buildings; another five were shot dead by the gates. To seal off any reinforcements, circling A-1s strafed a nearby vehicle bridge with white-phosphorous rockets. The main danger, however, seemed to be automatic-weapons fire from a two-story building in the center of the compound, which pinned down part of the force. A grenadier shot 40mm rifle-bombs into the doors and windows, obliterating the threat.

After ten minutes of searching, no prisoners had been found. The order was reluctantly given to withdraw, through the hole blown in the wall, to an extraction point 100 yards away. Power-poles were blown down to clear a landing zone. A third helicopter came in against enemy small-arms fire, which was soon suppressed, and the whole force departed the scene. On the way back some SAM sites were lit up and had to be evaded, but everyone arrived safely in Thailand. General Manor flew to Udorn to congratulate "a very disappointed group of people."

The mission was in many ways a success. Simons' raiders killed 25-40 enemy and got home without loss. The only casualty had been caused by a fire extinguisher, which fell off its pins and broke a soldier's ankle during the hard helicopter landing. But there was one minor snafu: the fourth helicopter had mistaken a nearby school for Sontay and landed in the wrong place, as DIA had warned might happen. And of course there was one not-so-minor mistake: the prisoners were not there.

Whose fault was that? Journalistic consensus blamed "the intelligence community" for "failing to verify" the presence of POWs at Sontay, for failing to provide the combat teams with "accurate information," and for an overall performance that was "woefully inadequate." But whatever else it was, Sontay was not an intelligence failure. DIA's map through the world's most sophisticated air-defense system had been accurate. Its warning about the school near Sontay had been right. Its prediction the camp would be empty was fulfilled. As in so many other supposed "intelligence failures," the problem was not a lack of facts. The problem was the facts were ignored by commanders, who refused to cancel a course of action to which they were already pledged.