"In God we trust, all others we monitor."
—Intercept operator's motto
NSA study, Deadly Transmissions,
December 1970
"The public has a duty to watch its Government closely and keep it on
the right track."
Lieutenant General Kenneth A.. Minihan,
USAF Director, National Security
Agency
NSA Newsletter,
June 1997
"The American people have to trust us and in order to trust us they have
to know about us."
Lieutenant General Michael V. Hayden,
USAF
Director, National Security Agency
Address on October 19, 2000
"Behind closed doors, there is no guarantee that the most basic of
individual freedoms will be preserved. And as we enter the 21st Century,
the great fear we have for our democracy is the enveloping culture of
government secrecy and the corresponding distrust of government that
follows."
Senators Daniel Patrick Moynihan
and Rob Wyden
U.S. Senate Report,
Secrecy in International and
Domestic Policy Making:
The Case for More Sunshine,
October 2000
CHAPTER ONE
MEMORY
KVZIEBCEN CKYIECDVG DBCOOVK HN CKYCFEUFJ ECZHIKUCF
MIBEVG FHOHFD NQXVWXIV NWQFWQG HG IMF FH EOF AB EWHB
XI GAEEXD WJP JZPWC ABCADL WP TYA RIW 'DYPJ YPWBOYS' XL
AXLB APYTIOWL ENTOJXGCM FVMMCD ND ENJBMD FGXMD VGXM
OG BMDO RPI EKFSKRPJV OXUVAZPJ QXSHJXSAVP HJXHXVKE LXJ
Z.Q. JPLXJSV
His step had an unusual urgency to it. Not fast, but anxious, like a child
heading out to recess who had been warned not to run. It was late
morning and the warm, still air had turned heavy with moisture, causing
others on the long hallway to walk with a slow shuffle, a sort of somber
march. In June 1930, the boxy, sprawling Munitions Building, near the
Washington Monument, was a study in monotony. Endless corridors
connecting to endless corridors. Walls a shade of green common to bad
cheese and fruit. Forests of oak desks separated down the middle by
rows of tall columns, like concrete redwoods, each with a number
designating a particular workspace.
Oddly, he made a sudden left turn into a nearly deserted wing. It was
lined with closed doors containing dim, opaque windows and empty
name holders. Where was he going, they wondered, attempting to keep
up with him as beads of perspiration wetted their brows. At thirty-eight
years old, the Russian-born William Frederick Friedman had spent most of his adult life studying, practicing, defining the black art of code
breaking. The year before, he had been appointed the chief and sole
employee of a secret new Army organization responsible for analyzing
and cracking foreign codes and ciphers. Now, at last, his one-man Signal
Intelligence Service actually had employees, three of them, who were
attempting to keep pace close behind.
Halfway down the hall Friedman turned right into Room 3416, small
office containing a massive black vault, the kind found in large banks.
Reaching into his inside coat pocket, he removed a small card. Then,
standing in front of the thick round combination dial to block the view,
he began twisting the dial back and forth. Seconds later he yanked up
the silver bolt and slowly pulled open the heavy door, only to reveal
another wall of steel behind it. This time he removed a key from his
trouser pocket and turned it in the lock, swinging aside the second door
to reveal an interior as dark as a midnight lunar eclipse.
Disappearing into the void, he drew out a small box of matches and lit
one. The gentle flame seemed to soften the hard lines of his face: the
bony cheeks; the pursed, pencil-thin lips; the narrow mustache, as
straight as a ruler; and the wisps of receding hair combed back tight
against his scalp. Standing outside the vault were his three young hires.
Now it was time to tell them the secret. Friedman yanked on the dangling
cord attached to an overhead light bulb, switched on a nearby fan to
circulate the hot, stale air, and invited them in. "Welcome, gentlemen,"
he said solemnly, "to the secret archives of the American Black
Chamber."
Until a few weeks before, none of the new recruits had had even the
slightest idea what codebreaking was. Frank B. Rowlett stood next to a
filing cabinet in full plumage: blue serge jacket, white pinstriped
trousers, and a virgin pair of white suede shoes. Beefy and round-faced,
with rimless glasses, he felt proud that he had luckily decided to wear
his new wardrobe on this day. A high school teacher from rural southern
Virginia, he received a degree in math the year earlier from Emory and
Henry College, a small Virginia school.
The two men standing near Rowlett were a vision of contrasts. Short,
bespectacled Abraham Sinkov; Brooklynite Solomon Kullback, tall and
husky. Both were high school teachers from New York, both were
graduates of City College in New York, and both had received master's
degrees from Columbia.
Like a sorcerer instructing his disciples on the mystic path to eternal
life, Friedman began his introduction into the shadowy history of
American cryptology. In hushed tones he told his young employees about
the Black Chamber, America's first civilian codebreaking organization.
How for a decade it operated in utmost secrecy from a brown-stone in New York City. How it skillfully decoded more than 10,000 messages
from nearly two dozen nations, including those in difficult Japanese
diplomatic code. How it played the key role in deciphering messages to
and from the delegates to the post-World War I disarmament talks, thus
giving the American delegation the inside track. He told of Herbert
Osborne Yardley, the Black Chamber's hard-drinking, poker-playing
chief, who had directed the Army's cryptanalytic activities during the
war.
Then he related the story of the Chamber's demise eight months
earlier. How the newly appointed secretary of state, Henry Stimson, had
become outraged and ordered its immediate closing when he discovered
that America was eavesdropping on friends as well as foes. Friedman told
of the firing of Yardley and the rest of the Chamber's employees and of
how the government had naively taken itself out of the code-breaking
business.
It was a troubling prospect. If a new war were to break out, the United
States would once again have to start from scratch. The advances
achieved against Japan's codes would be lost forever. Foreign nations
would gain great advantage while the United States clung to diplomatic
niceties. Standing in the vault containing the salvaged records of the old
Black Chamber, Friedman told his three assistants, fresh out of college,
that they were now the new Black Chamber. The Army, he said, had
given its cautious approval to secretly raise the organization from the
ashes, hide it deep within the bureaucracy, and rename it the Signal
Intelligence Service. The State Department, they were sternly warned,
was never to know of its existence.
In late June 1930, America's entire cryptologic body of secrets—
personnel, equipment and records—fit comfortably in a vault twenty-five
feet square.
On the southbound lane of the Baltimore—Washington Parkway, near
the sleepy Maryland hamlet of Annapolis Junction, a restricted, specially
constructed exit ramp disappears quickly from view. Hidden by tall
earthen berms and thick trees, the ramp leads to a labyrinth of barbed wire
fences, massive boulders placed close together, motion detectors,
hydraulic antitruck devices, and thick cement barriers. During alerts,
commandos dressed in black paramilitary uniforms, wearing special
headgear, and brandishing an assortment of weapons including Colt
9mm submachine guns, quickly respond. They are known as the "Men in
Black." Telephoto surveillance cameras peer down, armed police patrol
the boundaries, and bright yellow signs warn against taking any
photographs or making so much as a note or a simple sketch, under the
penalties of the Internal Security Act. What lies beyond is a strange and invisible city unlike any other on earth. It contains what is probably the
largest body of secrets ever created.
Seventy-one years after Friedman and his three new employees
gathered for the first time in their vault, with room to spare, the lineal
descendant of the Black Chamber now requires an entire city to contain
it. The land beyond the steel-and-cement no-man's-land is a dark and
mysterious place, virtually unknown to the outside world. It is made up
of more than sixty buildings: offices, warehouses, factories, laboratories,
and living quarters. It is a place where tens of thousands of people work
in absolute secrecy. Most will live and die without ever having told their
spouses exactly what they do. By the dawn of the year 2001, the Black
Chamber had become a black empire and the home to the National
Security Agency, the largest, most secret, and most advanced spy
organization on the planet.
Known to some as Crypto City, it is an odd and mysterious place,
where even the priests and ministers have security clearances far above
Top Secret, and religious services are held in an un-buggable room. "The
NSA Christmas party was a big secret," recalled one former deputy
director of the agency. "They held it at Cole field house but they called it
something else." Officials hold such titles as Chief of Anonymity, and
even the local newsletter, with its softball scores and schedules for the
Ceramic Grafters Club, warns that copies "should be destroyed as soon
as they have been read." Crypto City is home to the largest collection of
hyper powerful computers, advanced mathematicians, and language
experts on the planet. Within the fence, time is measured by the
femtosecond—one million billionth of a second—and scientists work in
secret to develop computers capable of performing more than one
septillion (1,000,000,000,000,000,000,000,000) operations every second.
Nearby residents can only guess what lies beyond the forbidden exit
ramp. County officials say they have no idea how many people work
there, and no one will tell them. Traffic planners from the county
planning department, it is said, once put a rubber traffic-counting cord
across a road leading to the city, but armed guards came out and quickly
sliced it. "For a long time we didn't tell anybody who we were," admitted
one agency official. "The focus was not on community activity. It was like everyone outside the agency was the enemy."
In an effort to ease relations with its neighbors, officials from the city
gave Maryland's transportation secretary, James Lighthizer, a rare tour.
But the state official was less than overwhelmed. "I didn't get to see a
darn thing," he said.
At a nearby gas station, owner Clifford Roop says the people traveling
into and out of the city keep to themselves. "They say they work for the
DoD [Department of Defense]. They don't talk about their work at all."
Once, when a reporter happened into the station and began taking a few
notes, two police cruisers from the secret city rushed up to the office and
demanded an ID from the journalist. This was not an unusual response.
When a photographer hired by real estate developers started up a hill
near Crypto City to snap some shots of a future construction site, he was
soon surrounded by NSA security vehicles. "They picked him up and
hauled him in and asked what he was doing," said Robert R. Strott, a
senior vice president at Constellation Real Estate, which was a partner in
the project. During interrogation the photographer not only denied
attempting to take a shot of Crypto City, he said he had never even heard
of NSA. Worried that occupants of an eleven-story office building might
be able to look into the city, NSA leased the entire building before it was
completed.
To dampen curiosity and keep peace with the neighbors, NSA director
William O. Studeman, a three-star admiral, once gave a quiet briefing to
a small group of community leaders in the area. "I do this with some
trepidation," he warned, "because it is the ethic of the agency—
sometimes called in the vernacular the super secret NSA—to keep a low
profile." Nevertheless, he gave his listeners a brief idea of NSA's
tremendous size. "We're the largest and most technical of all the U.S.
intelligence agencies. We're the largest in terms of people and we're the
largest in terms of budget. . . . We have people not only here at NSA but
there are actually more people out in the field that we have operational
control over—principally military—than exist here in Maryland. . . . The
people number in the tens of thousands and the budget to operate that
system is measured in the billions of dollars annually—billions
annually."
A decade ago, on the third floor of Operations Building 1 at the heart
of the sprawling city, a standing-room-only crowd packed a hall. On
stage was Frank Rowlett, in whose honor an annual award was being
established. As he looked out toward the audience in the Friedman
Auditorium, named after his former boss, his mind no doubt skipped
back in time, back to that hot, sticky, June afternoon in 1930 when he
walked into the dim vault, dressed in his white suede shoes and blue
serge jacket, and first learned the secrets of the Black Chamber. How big
that vault had grown, he must have marveled.
For most of the last half of the twentieth century, that burgeoning
growth had one singular objective: to break the stubborn Russian cipher
system and eavesdrop on that nation's most secret communications. But
long before the codebreakers moved into the sterile supercomputer
laboratories, clean rooms, and anechoic chambers, their hunt for the
solution to that ultimate puzzle took them to dark lake beds and through
muddy swamps in the early light of the new Cold War.
CHAPTER TWO
SWEAT
YNTS QHABT YBK KJVT NR ORLSJN HCTCYA HQYKJV CYOCMBYNT
GXRYK SXRKVWNRNIO YJVONHB NH VH KXASH OAXBBJNHB WNHB
KSXXMT, FVTS SVJYMBF CFI El BNSYYC JTMKEID
AXITUL PGGTXLW VGA OCXFT AUMCAL VAGH RXDKQPUR PXDM
HQRSESTYY TBDSPKTTY YTT ERYHURBRWCVRPW RW
JCBRSKJURTWESK DPSRHRTY
The wet, fertile loam swallowed the corporal's boots, oozing between the
tight laces like melted chocolate. The spring night was dark and cool and
he was walking backward in the muck, trying to balance his end of the
heavy box. More men followed, each weighted down with stiff crates that
gave off the sweet aroma of fresh pine. Except for the chirping sound of
crickets, and an occasional grunt, the only sounds to be heard were
sudden splashes as the heavy containers were tossed from boats into the
deepest part of the lake. Germany would keep its secrets.
It was the final night of April 1945. A few hundred miles away, in a
stale bunker beneath Berlin, Adolf Hitler and his new bride bid a last
farewell to each other, to the Reich, and to the dawn. The smoldering
embers of Nazism were at long last dying, only to be replaced by the
budding flames of Soviet Communism.
Just five days after Hitler's postnuptial suicide, General William O.
Donovan, chief of the Office of Strategic Services, delivered a secret
report to President Harry Truman outlining the dangers of this new
conflict. Upon the successful conclusion of World War II, Donovan
warned, "the United States will be confronted with a situation potentially
more dangerous than any preceding one." Russia, he cautioned, "would
become a menace more formidable to the United States than any yet
known."
For nearly a year both Washington and London had been secretly
planning the first battle of the new Cold War. This war, unlike the last,
would have to be fought in the shadows. The goal would be the capture
of signals rather than cities; complex mathematical algorithms and
whirring computers, rather than brawn and bullets, would determine the
winner. The work would be known as signals intelligence—"Sigint," to the
initiated—a polite term for "reading someone else's mail." Sigint would
include both communications intelligence (Comint), eavesdropping on
understandable language, and electronic intelligence (Elint), snatching
signals from such things as radar.
More than a month before Hitler's death, the battle began: a small
team of American and British codebreakers boarded airplanes and headed across the English Channel. The team was part of a unique,
highly secret organization with the cover name TICOM, short for Target
Intelligence Committee. Its mission, in the penultimate days of the war,
was to capture as many German codebreakers and cipher machines as
possible. With such information, Allied cryptologists could discover
which of their cipher systems might have been broken, and thus were
vulnerable to attack. At the same time, because the Germans had
developed advanced systems to attack Soviet codes and ciphers, the West
would gain an invaluable shortcut in finding ways to break Russian
cipher systems. The key, however, was finding the men and machines
before the Russians, who could then use the German successes to break
American and British ciphers.
Colonel George A. Bicher, the director of the U.S. Signal Intelligence
Division in Europe, conceived of TICOM in the summer of 1944. The
organization was so secret that even today, more than half a century
later, all details concerning its operations and activities remain classified
higher than Top Secret by both the American and British governments.
In 1992, the director of the National Security Agency extended the
secrecy order until the year 2012, making TICOM probably the last great
secret of the Second World War.
Senior commanders on both sides of the Atlantic quickly saw the
potential in such an organization. In August 1944, General George C.
Marshall, the U.S. Army Chief of Staff, sent a codeword radio message to
General Dwight D. Eisenhower at his headquarters in London instructing
him to give TICOM the highest priority. Later that day, he followed up
with a laundry list to Eisenhower detailing the items he wanted TICOM to
capture, including all the codemaking and code-breaking documents and
equipment they could get their hands on.
TICOM's members were among the few who knew the Ultra secret,
that the United States and Britain had broken Germany's highest-level
codes. And they knew that whoever won the race to Hitler's cache of
cryptologic secrets held the advantage in the next war, whether cold or
hot. Because many of the members of TICOM would go on to run both
NSA and the British postwar code breaking center, it was a war they
themselves would eventually have to fight.
For more than four years, the best German cryptanalysts had been
attacking American, British, and Russian code and cipher systems, with
deadly success. With luck, somewhere in the ruins the Allies would find
a key that could unlock a number of complex Soviet codes, saving years
of frustrating work. And some locked vault might also contain reams of
intercepted and decoded Russian messages, which would offer enormous
insight into Soviet military and political intentions after the war. At the
same time, the interrogation transcripts and other documents could shed
light on unknown weaknesses in American and British cryptography,weaknesses that might prove fatal in any future conflict.
Because all of the key cryptologic targets were located in Berlin, there
was added urgency: Russian forces would shortly occupy that area.
Thus, "the plan contemplated a simultaneous seizure and exploitation of
the chief Sigint centers through an air-borne action," said the TICOM
report. These centers had been pinpointed by means of Ultra decrypts:
messages that had been encrypted by Germany's high-level cipher
machine, the Enigma, and decoded by British and American
codebreakers.
As outlined in the TICOM reports, there were four principal objectives:
a. To learn the extent of the German cryptanalytic effort against
England and America;
b. To prevent the results of such German cryptanalysis against
England and America from falling into unauthorized hands as the
German Armies retreated;
c. To exploit German cryptologic techniques and inventions before
they could be destroyed by the Germans; and
d. To uncover items of signal intelligence value in prosecuting the
war against Japan.
"The TICOM mission was of highest importance," the document
concluded. "American cryptographers did not then know with certainty
the extent to which United States communications were secure or
insecure, nor did they know the extent of the enemy's cryptanalytic
abilities, strength, and material."
TICOMs plan to quickly snatch up the people, papers, and equipment
as the Nazi war machine began to collapse was nearly completed by
Christmas, 1944. But within months, Germany was in chaos; Hitler's
codebreaking agencies began to scatter. The original plan, said the
report, "was no longer feasible." The chances that Anglo-American
parachute teams might seize worthwhile personnel and material, and
then hold them through the final battles, became remote.
Instead, TICOM decided to alert six teams in England and send them
into enemy territory as United States and British troops were
overrunning it. The teams were to "take over and exploit known or newly
discovered targets of signal intelligence interest and to search for other
signal intelligence targets and personnel."
It was in drafty brick buildings on a drab Gothic-Victorian estate
called Bletchley Park that the future TICOM team members had labored
12
during much of the war. Hidden away in the foggy English county of
Buckinghamshire, Bletchley was formally known as the Government
Code and Cypher School. After the war it changed its name to the less
descriptive Government Communications Headquarters (GCHQ). The
suburban location was chosen because it was halfway between the
universities of Oxford and Cambridge, key locations for finding new
recruits, and only forty-seven miles from London.
In their Spartan offices the eclectic band of mathematicians, linguists,
and English professors molded their intellects into what was possibly the
deadliest weapon of the war against Germany. As the final TICOM report
makes clear, the German high-level cryptography "was brilliantly
conceived," but the cryptanalytic breakthroughs of the British and
American codebreakers were "more brilliantly conceived."
So good was the Allied ability to eavesdrop on a wide range of German
communications that it has recently led to troubling questions about
how early in the war the Allies discovered evidence of the Holocaust.
"Allied Comint agencies had been exploiting a number of French codes
and ciphers from the beginning of the war," NSA historian Robert J.
Hanyok recently told a gathering in the agency's Friedman Auditorium.
"They soon found reflections of the anti-Jewish laws in their intercept of
both Vichy diplomatic and colonial radio and cable traffic." Pressured by
the German occupation authorities, France in 1942 began rounding up
Jews for shipment to "resettlement sites," a euphemism for concentration
camps.
According to a comprehensive NSA study undertaken by Hanyok,
Allied communications intelligence would have picked up indications of
this roundup from the cable lines and airwaves linking Vichy France with
foreign capitals. The communications lines would have been buzzing with
pleas by worried relatives for information on loved ones interned in
various French camps. But in the end, Hanyok noted, only a fraction of
the intercepts were ever distributed and the principal focus was always
on strategic military traffic, not routine diplomatic communications.
"Intelligence on the Holocaust was NOT critical to Allied strategy," said
Hanyok [emphasis in original]. "Did Comint reveal the Holocaust, and,
especially, its final aim?" he asked. "The real problem," he concluded,
"was not interpreting the intelligence, but the attitude by the Allies, and
the rest of the world, that the unthinkable was actually happening."
In March 1945, as the damp chill of a long English winter began to
fade, TICOM teams began to fan out across Germany in search of
codebreakers and their books and equipment. "One day we got this
frantic call," said Paul E. Neff, a U.S. Army major assigned to Bletchley
Park. "They had run across these people, Germans, in this castle . . . who had been in the cryptographic business, signals intelligence, all of
them. Bongo. Quickly Bletchley sent me." Within a few days, Neff was at the castle, in the German state of Saxony.
"The war was still going on and we were pretty far forward," Neff said.
"We sorted the people out, interrogated, tried to find out what they were
working on, where they had stood with it, tried to get our hands on all
the papers that were left. . . . But my problem became, What are we
going to do with them? Because they apparently had a lot of good
information. . . . These Germans, as you might know, had been working
on the Russia problem too." Neff had stumbled into a gold mine, because
not only had the codebreakers worked on Russian codes and ciphers,
but the castle contained a German Foreign Office signals intelligence
archive. Neff's dilemma was the location of the castle, which was located
in territory assigned to the Soviets—and Russian troops were quickly
moving into the area. Neff needed to get the people and codebreaking
materials out fast.
Neff contacted Colonel George Bicher, in charge of the American
TICOM unit in London, and suggested shipping the documents—and the
German codebreakers—to England. But the issue of transporting the
prisoners across the English Channel became very sensitive. "Apparently
they had a hard time when this thing hit London because they couldn't
decide what to do. They had to clear it up to the attorney general or
whatever he's called over there. Is it legal to do?" Eventually the British
agreed to have the Germans secretly transferred to England. "We got a
plane one day," said Neff, "escorted this crowd down to the airfield, put
them on the plane, and flew them over to London. The British picked
them up over there and gave them a place to stay, fed them, and
interrogated the hell out of them. Now, what happened to those TICOM
records I don't know." Two days later, Russian troops overtook that same
area.
The May morning was as dark as black velvet when Paul K. Whitaker
opened his eyes at 4:45. Short and stout, with a thick crop of light brown
hair, the American Army first lieutenant slowly began to wake himself
up. For two years he had been assigned to Hut 3, the section of Bletchley
Park that specialized in translating and analyzing the decrypted Enigma
Army and Air Force messages.
At thirty-eight, Whitaker was considerably older than his fellow junior
officers. For more than a decade before joining the Army in 1942 he had
studied and taught German in the United States as well as in Germany
and Austria, receiving his doctorate from Ohio State. While doing
graduate work at the University of Munich in 1930 he often dined at a
popular nearby tavern, the Osteria Bavaria. There, at the stark wooden
tables, he would frequently see another regular customer enjoying the
Koniginpastete and the russische Eier. Seated nearby, always at the same round table and surrounded by friends and associates, was a quiet but
ambitious local politician by the name of Adolf Hitler.
The first dim rays of light illuminated a fresh spring snow, surprising
Whitaker as he stepped out of his quarters. Like dusting powder, the
snow lent a certain beauty to the tired estate, gently filling in the cracks
on the red brick walls and softening the dark blemishes caused by years
of chimney soot.
Rather than head for Hut 3, Whitaker went straight to the bus stop at
Bletchley Park. Also waiting there was First Lieutenant Selmer S.
Norland, who had traveled to England with Whitaker several years
earlier. Raised in northern Iowa, Norland had the solid, muscular
features of a farmer and a serious face with deep-set eyes. Before
entering the Army in 1942 he taught history and German in a local high
school for three years and now worked as a translator in Hut 3 with
Whitaker.
At precisely 6:00 A.M. the special bus arrived, coughing thick diesel
fumes and cutting neat brown lines in the virgin snow. About a dozen
officers and enlisted men, both British and American, climbed aboard.
Seated near Whitaker was another American Army officer shipped over
several years earlier, Arthur Levenson, a tall, lean mathematician from
New York who worked in Hut 6 as a cryptanalyst. Like Whitaker and
Norland, Levenson, who also doubled as the secretary of the Bletchley
Park Chess Club, had spent time working on code problems before his
transfer to England. In July 1943 Whitaker, Norland, Levenson, and
seven other cryptologic officers boarded the huge British liner Aquitania
as it set sail for Scotland. A few weeks later they became the first U.S.
Army codebreakers to be assigned to Bletchley.
A soldier in the sentry box snapped a salute as the heavy bus pulled
out through the park's intricate iron gate. Like cenobite monks leaving
their monastery for the first time, the newest TICOM team had little idea
what to expect. Since the Enigma project's beginning, British policy had
forbidden sending anyone who worked on it into combat areas. For years
the Bletchley staff had been closeted voyeurs, reading about the war
through newspapers and purloined messages.
The snow-covered fields began merging into an endless white
comforter as the bus hurried through the Midlands toward London.
Sitting near the window, Howard Campaign certainly felt the
excitement. As a young instructor at the University of Minnesota with a
Ph.D. in mathematics, he sent the Navy a homemade design for an
encryption device. Although Navy officials turned down the invention,
they did offer him a correspondence course in cryptanalysis, which he
passed. "I eventually got my commission and it was dated 5 December
1941," Campaign recalled. "So two days later the balloon went up and we were in the war."
Now as the bus pulled up to Croydon Air Field for the flight to Paris
on the first leg of their mission, Campaign was about to lead the hunt
for a mysterious German cipher machine nicknamed the Fish.
Although Bletchley Park had conquered the Enigma machine, the
Germans had managed to go one better. They developed a new and even
more secret cipher machine, the Geheimschreiber, or secret writer, which
was reserved for the very-highest-level messages, including those to and
from Hitler himself. German cryptographers called an early model
Swordfish. The Americans and British simply called them the Fish.
Unlike Enigma, the Fish were capable of automatically encrypting at one
end and decrypting at the other. Also, rather than the standard 26-letter
alphabet, the Fish used the 32-character Baudot code, which turned the
machine into a high-speed teleprinter.
TICOMs goal was to capture a working model intact and thus learn
exactly how the Germans built such a complex, sophisticated encryption
device. Especially, they needed to discover faster and better ways to
defeat such machines in the future should they be copied and used by
the Russians.
The Royal Air Force flight to Paris was mostly smooth, reminding Paul
Whitaker of sailing in a boat through gentle swells. Along with a number
of the other men on the flight, he was on a plane for the first time. "The
impressions were amazingly lacking in strangeness," he jotted in his
small black notebook, "probably because one sees so many films taken
from aircraft. It seemed completely normal to be looking down on tiny
houses and fields a mile below."
Within a few days the team, packed into an olive-green, 2 ½ -ton U.S.
Army truck and an open jeep, pushed into Germany. Their target was a
suspected major Air Force signals intelligence center in the southern
Bavarian city of Kaufbeuren, a market center of medieval towers and
crumbling fortifications on the Wertach River. Fresh from their secret
monastery in the English countryside, many on the TICOM team were
unprepared for the devastation they witnessed. "The roads were lined
with burned-out and shot up tanks and vehicles of all sorts," Whitaker
jotted in his journal as he bounced along the road from Heidelberg, "and
many villages, even small ones, were badly smashed up and burned."
Around midnight, they arrived at Augsburg, a city that would soon
become one of NSA's most secret and important listening posts in
Europe. The next morning, having spent the night in a former German
Air Force headquarters, the team discovered a communications center in
the basement. In some of these buildings the Allies had moved in so fast
that the ghosts of the former occupants still seemed to be present. The
Germans had departed with such haste from one facility that when the Americans arrived the teleprinters were still disgorging long thin message
tapes.
Other teleprinters provided insight into the private horror of defeat.
"How are things down there?" read one tape still dangling from the
machine. Whitaker saw it was from a soldier in the cathedral town of
Ulm to a colleague in Augsburg. "Reports here say that the Americans
are in Augsburg already." "No," the soldier replied, "everything here is
O.K." But suddenly he added, "My God, here they are, auf Wiedersehen."
Within a few days the team struck gold. They came upon an entire
convoy of four German signal trucks, complete with four Fish machines,
a signals technician, German drivers, and a lieutenant in charge. Arthur
Levenson and Major Ralph Tester, a British expert on the Fish, escorted
the whole lot, including the Germans, back to England. Once at Bletchley
Park the machines were reverse-engineered to determine exactly how
they were built and how they operated. (Levenson would later return to
Washington and go on to become chief of the Russian code-breaking
section at NSA.)
With enough Fish and other equipment to keep the engineers busy for
a long time at Bletchley, the team began a manhunt for key German
codebreakers. On May 21, 1945, Lieutenant Commander Howard
Campaigne and several other TICOM officers interviewed a small group of
Sigint personnel being held in Rosenheim. They had all worked for a unit
of the Signals Intelligence Agency of the German Abwehr High Command,
a major target of TICOM. What the prisoners told Campaign would lead
to one of the most important, and most secret, discoveries in the history
of Cold War code breaking. Their command, they said, had built a
machine that broke the highest-level Russian cipher system. The
machine, now buried beneath the cobblestones in front of a building
nearby, had been designed to attack the advanced Russian teleprinter
cipher—the Soviet equivalent of the Fish.
If this was true, it was breathtaking. For over six years U.S. and
British codebreakers had placed Japan and Germany under a
microscope, to the near exclusion of Russia and almost all other areas.
Now with the war over and with Communist Russia as their new major
adversary, the codebreakers would have to start all over from scratch.
Rut if a working machine capable of breaking high-level Russian ciphers
was indeed buried nearby, years of mind-numbing effort would be saved.
The Germans, eager to be released from prison, quickly agreed to lead
TICOM to the machine. Campaign wasted no time and the next day the
twenty-eight prisoners, dressed in their German Army uniforms, began
pulling up the cobblestones and opening the ground with picks and
shovels. Slowly the heavy wooden boxes began to appear. One after
another they were pulled from the earth, until the crates nearly filled the grounds. In all there were a dozen huge chests weighing more than 600
pounds each; 53 chests weighing nearly 100 pounds each; and about 53
more weighing 50 pounds each. It was a massive haul of some 7 ½ tons.
Over the next several days the dark gray equipment was carefully lifted from its crates and set up in the basement of the building. Then, like magic, high-level encrypted Russian communications, pulled from the ether, began spewing forth in readable plaintext. Whitaker, who pulled into the camp a short time later, was amazed. "They were working like beavers before we ever arrived," he scribbled in his notebook. "They had one of the machines all set up and receiving traffic when we got there."
The Russian system involved dividing the transmissions into nine separate parts and then transmitting them on nine different channels. The German machines were able to take the intercepted signals and stitch them back together again in the proper order. For Campaigne and the rest of the TICOM team, it was a once-in-a-lifetime discovery. Back in Washington, Campaigne would eventually go on to become chief of research at NSA.
Once the demonstration was over, Campaigne had the German soldiers repack the equipment and the next day it was loaded on a convoy, completely filling four heavy trucks. Two TICOM members, including First Lieutenant Selmer Norland, who would also go on to a long career at NSA, accompanied the equipment and soldiers back to England. There it was set up near Bletchley Park and quickly put into operation. It, or a working model, was later shipped back to Washington. The discovery of the Russian code breaking machine was a principal reason why both the U.S. and British governments still have an absolute ban on all details surrounding the TICOM operations.
All told, the TICOM teams salvaged approximately five tons of German Sigint documents. In addition, many cryptologic devices and machines were found and returned to Bletchley.
Equally important were the interrogations of the nearly 200 key German codebreakers, some of which were conducted at a secret location codenamed Dustbin. In addition to the discovery of the Russian Fish, another reason for the enormous secrecy surrounding TICOM may be the question of what happened to the hundreds of former Nazi code-breakers secretly brought to England. Were any of the war criminals given new identities and employed by the British or American government to work on Russian code breaking problems? Among those clandestinely brought into the United States was the top codebreaker Dr. Erich Huettenhain. "It is almost certain that no major cryptanalytic successes were achieved without his knowledge," said one TICOM document.
Among the surprises to come out of the interrogations was the fact that the Germans knew all along that Enigma was not totally secure. "We found that the Germans were well aware of the way the Enigma could be broken," recalled Howard Campaigne. "But they had concluded that it would take a whole building full of equipment to do it. And that's what we had. A building full of equipment. Which they hadn't pictured as really feasible."
In Washington, the TICOM materials were of enormous help in determining just how secure, or insecure, America's own cryptographic systems were. The picture painted by the documents and interrogations showed that while a number of lower-level systems had been read by German codebreakers, the most important ciphers remained impenetrable. "European cryptanalysts were unable to read any U.S. Army or Navy high-level cryptographic systems," the highly secret report said.
The Germans were never able to touch America's "Fish," a machine known as the SIGABA. Like the Fish, SIGABA was used for the Army and Navy's most sensitive communications. In fact, because TICOM showed that the SIGABA survived the war untouched by enemy codebreakers, it remained in service for some time afterward. It was finally taken out of service only because it did not meet the speed requirements of modern communications.
The TICOM report also indicated that other systems were not secure. One Army system and one Navy system were read for a short time. Both of the un-enciphered War Department telegraph codes were read by the Germans, and Hungary received photostats of War Department Confidential Code Number 2, probably from the Bulgarians. Also, thanks to a spy, Military Intelligence Code Number 11, which was used by the military attaché in Cairo, was read throughout the summer of 1942.
The most serious break was the solving of the Combined Naval Cypher Number 3, used by U.S. and Royal Navy convoy operations in the Atlantic; this Axis success led to many deaths. Other systems were also broken, but they were of less importance than the Allied breaks of Enigma and Fish.
By far the greatest value of TICOM, however, was not in looking back but in looking forward. With the end of the war, targets began shifting, the signals intelligence agencies dramatically downsized, and money became short. But at the start of the Cold War, as a result of TICOM, America had a significant lead. Not only did the U.S. code-breakers now have a secret skeleton key to Russia's Fish machine, it had a trapdoor into scores of code and cipher systems in dozens of countries. As a result of the German material and help from the British, for example, diplomatic communications to and from Afghanistan became "practically 100% readable." Thus, when Soviet officials discussed Asian diplomatic issues with the Afghan prime minister, the U.S. could listen in.
It was a remarkable accomplishment. At the outbreak of the war in Europe in 1939, the United States was attacking the systems of only Japan, Germany, Italy, and Mexico. But by the day the war ended, according to the TICOM report, "cryptanalytic attack had been directed against the cryptographic systems of every government that uses them except only our two allies, the British and the Soviet Union." Now readable, either fully or partially, were the encryption systems of Argentina, Belgium, Brazil, Bulgaria, Chile, China, Colombia, the Dominican Republic, Egypt, Ecuador, Ethiopia, Finland, France, Greece, Hungary, Iran, Iraq, Ireland, Italy, Japan, Lebanon, Mexico, the Netherlands, Peru, Portugal, Saudi Arabia, Spain, Switzerland, Syria, Thailand, Transjordan, Turkey, Uruguay, Venezuela, and Yugoslavia.
Between the attack on Pearl Harbor and August 1945, the Army's Signal Security Agency's Language Branch scanned more than 1 million decrypted messages and, of those, forwarded approximately 415,000 translations. But then it was over. Brigadier General W. Preston Corderman, chief of the Army codebreakers, was sure there would no longer be a need for much of a cryptanalytic effort. He therefore assembled the staff beneath the tall maple trees that gave his headquarters shade in the summer. The war was over, he told them, and so was their country's need for their services.
"Overnight, the targets that occupied most of the wartime cryptologic resources—Germany and Japan—had become cryptologic nonentities," said one NSA report. "One by one the radio receivers that had been faithfully tuned to enemy signals were switched off. Antenna fields were dismantled, equipment mothballed as station after station around the world ceased monitoring the airwaves, turned off the lights and padlocked the doors. Gone were the Army intercept stations at Miami, Florida; at New Delhi, India; at OSS Operations in Bellmore, New York; at Tarzana, California; and at Accra on the African Gold Coast. Silent were the Radio Intelligence Companies supporting General MacArthur in the Southwest Pacific and the Signal Service Companies in Europe."
The relative handful of American codebreakers who stayed on quickly shifted gears. The Soviet Union instantly became their number one target.
One key listening post not shut down was Vint Hill Farms Station. Known as Monitoring Station Number 1, it was located in the rural Virginia town of Warrenton. During the war, Vint Hill played a pivotal role in eavesdropping on enemy communications for thousands of miles in all directions. At war's end, 2,600 people stayed on, many of them intercept operators, to handle the transition from hot war to cold war.
They were able to eavesdrop on key Russian diplomatic and military communications sent over the Fish machine. "They intercepted printers at Vint Hill, Russian printers," said Colonel Russell H. Horton, who commanded the station shortly after the end of the war. "They had these ... circuits that had nine channels if I'm not mistaken. They had machines all hooked up so that they separated the channels and did all of the interception in Cyrillic characters." Horton added, "As far as I know, there was no effort against the Russians until after the war."
Although the fact was known to only a few, a small group of codebreakers had in fact been working on Russian code problems during the war. In 1943, American intelligence began to worry about a possible alliance between Nazi Germany and Russia as part of a comprehensive peace deal. Such a merger would have been a nightmare for the Allies. As a result, a few Army cryptanalysts were pulled away from work on German systems and assigned to a highly secret new unit with the goal of attempting to solve the enormously complex Soviet codes and ciphers.
Since 1939, thousands of encrypted Soviet messages, sent between Moscow and Washington, had been acquired from Western Union and other commercial telegraph companies. A major break occurred when it was discovered that identical code groups turned up in seven pairs of messages. To find even a single pair was a billion-to-one shot. Army codebreakers had discovered a "bust," an error or anomaly that opens a crack into the cipher system. Such a bust might be caused, for example, by a malfunction in a random-number generator. This bust, however, was caused by the Soviets reusing pages from one-time pads—the violation of a cardinal cryptographic rule. One-time pads had become two-time pads. Cecil Phillips, a former senior NSA official, played a key role in the early Soviet-watching program. "For a few months in early 1942," he said, "a time of great strain on the Soviet regime, the KGB's cryptographic center in the Soviet Union for some unknown reason printed duplicate copies of the 'key' on more than 35,000 pages . . . and then assembled and bound these one-time pads. . . . Thus, two sets of the ostensibly unique one-time pad page sets were manufactured."
The decision by the Soviet codemakers to duplicate the pages was likely the result of a sudden shortage of one-time pads, a result of Hitler's invasion of Russia in June 1941. To quickly fill the enormous demand for the pads, Russian cryptographers likely chose the easiest course: carbon paper. Suddenly production was doubled while, it was reasoned, security was diminished only slightly.
Phillips estimated that between 1942 and 1948, when the last onetime pad was used, more than 1.5 million messages were transmitted to Soviet trade and diplomatic posts around the world. Of those, American codebreakers obtained about a million, 30,000 of which had been enciphered with the duplicate pages. But despite the bust, days and weeks of frustrating work were required to squeeze out a clear-text message from a cipher text. Even then, usually the most they would have was a long, out-of-date message concerning such things as shipping schedules of the Soviet Purchasing Commission.
For more than thirty years the codebreakers worked on those messages. By the time the file drawer was closed for the last time, in 1980, they had managed to read portions of more than 2,900 Soviet diplomatic telegrams sent between 1940 and 1948. Codenamed Venona, the program was one of the most successful in NSA's history. It played a major role in breaking up key Soviet espionage networks in the United States during the postwar period, including networks aimed at the secrets of the atomic bomb.
On April 25, 1945, as TICOM officers began sloshing through the cold mud of Europe, attempting to reconstruct the past, another group of codebreakers was focused on a glittering party half the earth away, attempting to alter the future.
Long black limousines, like packs of panthers, raced up and down the steep San Francisco hills from one event to another. Flower trucks unloaded roses by the bushel. Flashbulbs exploded and champagne flowed like water under the Golden Gate. The event had all the sparkle and excitement of a Broadway show, as well it should have. The man producing it was the noted New York designer Jo Mielziner, responsible for some of the grandest theatrical musicals on the Great White Way. "Welcome United Nations," proclaimed the bright neon marquee of a downtown cinema. The scene was more suited to a Hollywood movie premiere than a solemn diplomatic event. Crowds of sightseers pushed against police lines, hoping for a brief glimpse of someone famous, as delegates from more than fifty countries crowded into the San Francisco Opera House to negotiate a framework for a new world order.
But the American delegates had a secret weapon. Like cheats at a poker game, they were peeking at their opponents' hands. Roosevelt fought hard for the United States to host the opening session; it seemed a magnanimous gesture to most of the delegates. But the real reason was to better enable the United States to eavesdrop on its guests.
Coded messages between the foreign delegations and their distant capitals passed through U.S. telegraph lines in San Francisco. With wartime censorship laws still in effect, Western Union and the other commercial telegraph companies were required to pass on both coded and uncoded telegrams to U.S. Army codebreakers.
Once the signals were captured, a specially designed time-delay device activated to allow recorders to be switched on. Devices were also developed to divert a single signal to several receivers. The intercepts were then forwarded to Arlington Hall, headquarters of the Army codebreakers, over forty-six special secure teletype lines. By the summer of 1945 the average number of daily messages had grown to 289,802, from only 46,865 in February 1943. The same soldiers who only a few weeks earlier had been deciphering German battle plans were now unraveling the codes and ciphers wound tightly around Argentine negotiating points.
During the San Francisco Conference, for example, American codebreakers were reading messages sent to and from the French delegation, which was using the Hagelin M-209, a complex six-wheel cipher machine broken by the Army Security Agency during the war. The decrypts revealed how desperate France had become to maintain its image as a major world power after the war. On April 29, for example, Fouques Duparc, the secretary general of the French delegation, complained in an encrypted note to General Charles de Gaulle in Paris that France was not chosen to be one of the "inviting powers" to the conference. "Our inclusion among the sponsoring powers," he wrote, "would have signified, in the eyes of all, our return to our traditional place in the world."
In charge of the San Francisco eavesdropping and codebreaking operation was Lieutenant Colonel Frank B. Rowlett, the protégé of William F. Friedman. Rowlett was relieved when the conference finally ended, and he considered it a great success. "Pressure of work due to the San Francisco Conference has at last abated," he wrote, "and the 24- hour day has been shortened. The feeling in the Branch is that the success of the Conference may owe a great deal to its contribution."
The San Francisco Conference served as an important demonstration of the usefulness of peacetime signals intelligence. Impressive was not just the volume of messages intercepted but also the wide range of countries whose secrets could be read. Messages from Colombia provided details on quiet disagreements between Russia and its satellite nations as well as on "Russia's prejudice toward the Latin American countries." Spanish decrypts indicated that their diplomats in San Francisco were warned to oppose a number of Russian moves: "Red maneuver . . . must be stopped at once," said one. A Czechoslovakian message indicated that nation's opposition to the admission of Argentina to the UN.
From the very moment of its birth, the United Nations was a microcosm of East-West spying. Just as with the founding conference, the United States pushed hard to locate the organization on American soil, largely to accommodate the eavesdroppers and codebreakers of NSA and its predecessors. The Russians, on the other hand, were also happy to have the UN on American soil—it gave them a reason to ship dozens of additional spies across U.S. borders.
Since the discovery of the Russian Fish machine by TICOM at the end of the war, and the ability to read a variety of diplomatic, KGB, and trade messages as a result of the Venona breakthrough on Soviet one time pads, American codebreakers had been astonishingly lucky. Virtually overnight they were placed in what NSA has called "a situation that compared favorably to the successes of World War II." For several years, American codebreakers were able to read encrypted Soviet armed forces, police, and industry communications and the agency could put together "a remarkably complete picture of the Soviet national security posture." But then, almost overnight in 1948, everything went silent. "In rapid succession, every one of these cipher systems went dark," said a recent NSA report, which called it "perhaps the most significant intelligence loss in U.S. history." It forever became known at NSA as Black Friday.
Just as the United States had successfully penetrated secret Soviet
communications networks, so the Russians had secretly penetrated the
Army Security Agency and later the Armed Forces Security Agency
(AFSA), into which ASA had been folded. Although he was never charged
with espionage, a gregarious Russian linguist by the name of William
Weisband became the chief suspect. Born to Russian parents in Egypt in
1908, Weisband emigrated to the United States in the 1920s and became
a U.S. citizen in 1938. Four years later he joined the Signal Security
Agency and was assigned to Sigint activities in North Africa and Italy,
before returning to Arlington Hall and joining its Russian Section.
Although Weisband was not a cryptanalyst, his fluency in Russian gave
him unique access to much of what the Russian codebreakers were
doing. In 1950, after being suspended from work on suspicion of
disloyalty, he skipped a federal grand jury hearing on Communist Party
activity and, as a result, was convicted of contempt and sentenced to a
year in prison. He died suddenly of natural causes in 1967, always
having denied any involvement in espionage.
For American codebreakers, the lights could not have gone out at a worse time. In late June 1950, North Korean forces poured across the 38th Parallel into the south, launching the Korean War. Once again, as with Pearl Harbor, America was caught by surprise.
A year before the attack, the Army, Navy, and Air Force code-breaking organizations had been combined into a single unit, AFSA. But instead of establishing a strong, centralized organization to manage the growing worldwide signals intelligence operations, each service was allowed to retain control of both intercept and code breaking activities. That left little for the director of AFSA to direct. Nor could he even issue assignments to field units. They would first have to pass through each of the services, which could then accept them, change them, or simply ignore them. Herbert L. Conley, who was in charge of Russian traffic analysis at AFSA in the late forties, and later headed up Russian code-breaking at NSA, likened the organization to a "three-headed monster." "He couldn't control anything outside of the buildings that were occupied," he said of the director.
In the weeks leading up to the attack, Korea barely registered as a Sigint target for AFSA. Out of two priority lists, North Korea was number fifteen on the secondary list. From listening posts at Kamiseya, Japan, and several other locations, most of the intercept activity was directed at Russia. Communist China was also a high priority, with eighty-seven intercept operators and analysts focused on it. But because AFSA had not broken any important Chinese cipher systems, most personnel concentrated on traffic analysis, the examination of the message's "external indicators," such as its date and "to" and "from" lines. North Korea, on the other hand, was targeted by just two intercept operators at the time the war broke out. In all, they had collected a paltry two hundred messages, and none of those had been processed. "AFSA had no Korean linguists, no Korean dictionaries, no traffic analytic aids, and no Korean typewriters," said a later NSA analysis.
Despite the limited resources, clues were there. Buried in stacks of intercepted Soviet traffic as far back as February were messages pointing to large shipments of medical supplies going from Russia to Korea. Other messages, about the same time, revealed a sudden and dramatic switch toward targets in South Korea by Soviet radio direction-finding units.
Suddenly, at 3:30 on the morning of June 25, 1950, Joseph Darrigo, a U.S. Army captain and the only American on the 38th Parallel, was jarred awake by the teeth-rattling roar of artillery fire. At that moment North Korean ground forces, led by 150 Soviet T-34 tanks, began their massive push into South Korea. Darrigo managed to escape just ahead of the advancing troops and spread the alarm. "AFSA (along with everyone else) was looking the other way when the war started," said a recent, highly secret NSA review. The first word to reach Washington came from a news account by a reporter in Seoul.
Within days, the North Korean Army had captured Seoul and continued to steamroll south, seeking to unify the peninsula under the flag of communism. In response, American troops were quickly dispatched to provide assistance to South Korea as part of a United Nations force. By the end of the first week, 40,000 South Korean soldiers had been killed, captured, or declared missing in action.
Following the attack, AFSA began a quick push to beef up its ranks. The number of intercept positions targeting North Korean traffic jumped from two to twelve. Any signals even remotely North Korean were transmitted back to AFSA headquarters in Washington, arriving ten to twelve hours after intercept. Soon, new messages were arriving hourly and lights were burning around the clock.
Nevertheless, cryptanalysis was virtually nonexistent. In fact, the first few decrypts of enciphered North Korean air traffic were produced not by professional codebreakers but by an uncleared U.S. Army chaplain using captured codebooks. Seconded into Sigint duty, Father Harold Henry had spent a number of years in Korea, where he learned the language. Most analysts instead concentrated on traffic analysis and plaintext intercepts—highly useful because of poor communications security by the North Koreans during the early part of the war. Among the messages sent in the clear were secret battle plans.
Adding to the problems, it was three months before a small advanced Sigint unit actually arrived on the Korean peninsula. Radio direction finding was greatly hampered by the mountainous terrain. Then there were the supply shortages, outmoded gear, difficulties in determining good intercept sites, equipment ill-suited to frequent movement over rough terrain, and a significant lack of translators.
From the beginning, the ground war went badly. By the end of July, the Eighth Army, led by General Walton H. Walker, had been forced into a boxlike area known as the Pusan Perimeter, so named because it surrounded the southeastern port of Pusan. "When we got into the . . . Perimeter, you never saw a more beat-up bunch of soldiers," recalled former PFC Leonard Korgie. "The North Koreans had hellish numbers and equipment. We were very, very thin in both."
Walker's one advantage was a constant supply of Sigint, which provided him with such vital information as the exact locations of North Korean positions. Armed with this intelligence, he was able to maximize his limited men and resources by constantly moving them to where new attacks were planned. Finally, following MacArthur's daring amphibious landing at Inchon, a port located behind enemy lines, Walker's men broke out of their box and joined in the attack, putting North Korea on the defensive.
In one sense, Sigint in Korea was like a scene from Back to the Future. After planting a number of sound-detecting devices forward of their bunkers to give warning of approaching troops, ASA soldiers discovered that the devices also picked up telephone calls. So they began using them for intercept—a practice common during World War I but long forgotten. This "ground-return intercept," using the principle of induction, enabled the ASA to collect some Chinese and Korean telephone traffic. The downside, however, was that in order to pick up the signals the intercept operator had to get much closer to enemy lines than normal, sometimes as close as thirty-five yards.
"One of our problems in Korea was linguists, there were so few," said Paul Odonovich, an NSA official who served in Korea with the Army 26 Security Agency. Odonovich commanded a company of intercept operators on the front lines. Sitting in antenna-bedecked vans, they would mostly eavesdrop on North Korean "voice Morse," an unusual procedure whereby the North Korean military would read the Morse code over the communications channels rather than tap it out with a key. "They used the singsong 'dit-dot-dit-dit' business," said Odonovich.
Other units conducting low-level voice intercept (LLVI), as it was known, operated out of jeeps and bunkers close to the front lines. The intelligence was then disseminated directly to combat units. By the end of the war, twenty-two LLVI teams were in operation. Air Force intercept operators also had some successes. Operating from small islands off North Korea, Sigint units were able to intercept North Korean, Chinese, and Soviet instructions to their pilots. The intercept operators would then disguise the intelligence as "radar plots" and pass them on in near real time to U.S. pilots operating over North Korean territory. Once they received the information, their "kill ratio" increased significantly.
After the battle began, the most important question was whether China would intervene. Since the end of World War II, Army Sigint specialists had engaged in a haphazard attack on Chinese communications. In 1945, General George Marshall attempted to bring Nationalist leader Chiang Kai-shek and Communist boss Mao Tse-tung to the negotiating table. At Marshall's request, a small group of intercept operators eavesdropped on both sides during the talks.
But the operation was less than a success. A team set up in Nanjing to intercept Nationalist communications was hampered by unreliable electrical power. Another, which targeted Communist links from a listening post in Seoul, was plagued with "poor hearability." Ironically, as the United States struggled, the British had been secretly listening to Chinese Communist communications for years. From 1943 until 1947, the Government Code and Cypher School successfully monitored a link between Moscow and Mao's headquarters in Yan'an, China. But because the link was part of a clandestine Soviet network, the decision was made to keep the Americans in the dark until March 1946.
Nevertheless, from the messages that the United States was able to intercept, it was clear that the two groups preferred to settle their differences on the battlefield rather than at the conference table. As a result, the Marshall mission was withdrawn in 1946. Thereafter, ASA dropped its study of Chinese Communist military ciphers and communications and turned its attention almost exclusively toward Russia. It would prove a serious mistake. Three years later, in 1949, Mao triumphed and Chiang fled to the island of Formosa.
About the same time, a small team of Chinese linguists led by Milton Zaslow began eavesdropping on and analyzing Chinese civilian communications—private telephone calls and telegrams. Unencrypted government messages would also travel over these lines. Beginning in early summer 1950, AFSA began developing "clear and convincing evidence" that Chinese troops were massing north of the Yalu River.
In May and June, Sigint reports noted that some 70,000 Chinese troops were moving down the Yangtze River in ships toward the city of Wuhan. The next month a message intercepted from Shanghai indicated that General Lin Piao, the commander of Chinese army forces, would intervene in Korea. Later reports noted that rail hubs in central China were jammed with soldiers on their way to Manchuria. By September, AFSA had identified six field armies in Manchuria, near the Korean border, and ferries on the Yalu River were being reserved for military use.
All of these reports were fully available to the Joint Chiefs of Staff, the White House, and to General Douglas MacArthur, the commander of the UN forces. Nevertheless, when asked by President Truman on October 15 about the chances of Chinese intervention, MacArthur replied, "Very little."
The indications continued. On October 21, AFSA issued a Sigint report stating that twenty troop trains were heading toward Manchuria from Shanghai. Then, on November 7, AFSA intercepted a radiotelephone call made by an East European in Beijing. He reported that orders had been issued allowing every Chinese soldier to volunteer to fight in Korea, saying, "We are already at war here." That same month, intercept operators picked up an unencrypted order for 30,000 maps of Korea to be sent from Shanghai to the forces in Manchuria.
Finally, intercepts during the first three weeks of November revealed that Beijing was in a state of emergency, with authorities sponsoring mass demonstrations demanding intervention, imposing more stringent censorship, improving air defense, and commanding that any soldier or officer could volunteer to serve in Korea. A medical headquarters urgently ordered troops in Manchuria to receive immunizations for diseases that were prevalent in North Korea—smallpox, cholera, and typhoid fever. AFSA reports demonstrated clearly that the Chinese were making extensive preparations for war.
But despite the many Sigint clues, U.S. and South Korean forces were once again caught by surprise. Early on the bitter-cold morning of November 26, with trumpets braying, thirty Chinese divisions surged across the North Korean border and forced U.S. and South Korean armies to make a precipitous retreat southward, costing the lives of many American soldiers.
"No one who received Comint product, including MacArthur's own G-2 [intelligence chief] in Tokyo, should have been surprised by the PRC intervention in the Korean War," said a recent, highly classified NSA review. The review then pointed a finger of blame for the disaster directly at MacArthur. "During the Second World War, MacArthur had disregarded Comint that contradicted his plans," it said. "MacArthur's zeal [to press ahead] to the Yalu probably caused him to minimize the Comint indicators of massive PRC intervention just as he had earlier minimized 'inconvenient' Comint reports about the Japanese. He thus drove his command to great defeat in Korea."
By mid-1951, with the 38th Parallel roughly dividing the two sides, ASA headquarters was established in the western suburbs of Seoul, on the campus of Ewha College, the largest women's school in Asia. There, traffic analysts put together a nearly complete Chinese army order of battle. Also, when truce negotiations began in July 1951, ASA units eavesdropped on meetings among the North Korean negotiating team. But that same month, the earphones of most of the intercept operators went silent as the North Koreans switched much of their radio communications to the security of landlines. NSA later attributed this caution to secrets allegedly passed to the Russians by former AFSA employee William Weisband.
Toward the end of the war, there were some tactical successes. By 1952, AFSA had broken a number of Chinese cipher systems. "The . . . last three major pushes that the Chinese had against us, we got those lock, stock, and barrel, cold," recalled Odonovich. "So that when the Chinese made their advances on our positions they were dead ducks . . . we had the code broken and everything."
But critical high-level communication between and among the Chinese and North Koreans was beyond the AFSA codebreakers' reach. Gone was the well-oiled machine that had helped win World War II. In its place was a confusing assortment of special-interest groups, each looking upon the other as the enemy; no one had the power to bring them together. "It has become apparent," complained General James Van Fleet, commander of the U.S. Eighth Army in June 1952, "that during the between-wars interim we have lost, through neglect, disinterest and possibly jealousy, much of the effectiveness in intelligence work that we acquired so painfully in World War II. Today, our intelligence operations in Korea have not yet approached the standards that we reached in the final year of the last war." A year later NSA director Ralph Canine, an Army lieutenant general, concurred with Van Fleet's observation.
So bad was the situation that in December 1951 the director of the CIA, Walter Bedell Smith, brought the problem to the attention of the National Security Council. In his memorandum, Smith warned that he was "gravely concerned as to the security and effectiveness with which the Communications Intelligence activities of the Government are being conducted." He complained that American Sigint had become "ineffective," as a result of the "system of divided authorities and multiple responsibilities."
Smith then discreetly referred to the mammoth security breach, blamed on Weisband, that had led the Soviets to change their systems. "In recent years," he said, "a number of losses have occurred which it is difficult to attribute to coincidence." To preserve what he called "this invaluable intelligence source"—Sigint—Smith called on Truman to ask Secretary of Defense Robert A. Lovett and Secretary of State Dean G. Acheson to conduct a "thorough investigation" of the agency. Three days later, on December 13, 1951, Truman ordered the investigation.
Appointed to head the probe was George Abbott Brownell, a fiftythree-year-old New York attorney and former special assistant to the secretary of the Air Force. Over six months, Brownell and his committee of distinguished citizens took AFSA apart and put it together again. In the end, they viewed AFSA as a "step backward." By June 13, 1952, when he turned his report over to Lovett and Acheson, Brownell had a blueprint for a strong, centralized new agency with a director more akin to a czar than to the wrestling referee the post resembled. Both secretaries approved and welcomed the independent review and set about carrying out its recommendations.
Four months later on October 24, Lovett, David K. Bruce from the State Department, and Everett Gleason of the NSC entered the Oval Office for a 3:30 off-the-record meeting with the president. There, Truman issued a highly secret order scrapping AFSA and creating in its place a new agency to be largely hidden from Congress, the public, and the world. Early on the morning of November 4, as Truman was leaving a voting booth in Independence, Missouri, the National Security Agency came to life. But few gave the new agency much hope. "The 'smart money' was betting that the new organization would not last much longer than AFSA," scoffed one official.
That night, Dwight David Eisenhower was elected the thirty-fourth president of the United States.
next.... NERVES
Over the next several days the dark gray equipment was carefully lifted from its crates and set up in the basement of the building. Then, like magic, high-level encrypted Russian communications, pulled from the ether, began spewing forth in readable plaintext. Whitaker, who pulled into the camp a short time later, was amazed. "They were working like beavers before we ever arrived," he scribbled in his notebook. "They had one of the machines all set up and receiving traffic when we got there."
The Russian system involved dividing the transmissions into nine separate parts and then transmitting them on nine different channels. The German machines were able to take the intercepted signals and stitch them back together again in the proper order. For Campaigne and the rest of the TICOM team, it was a once-in-a-lifetime discovery. Back in Washington, Campaigne would eventually go on to become chief of research at NSA.
Once the demonstration was over, Campaigne had the German soldiers repack the equipment and the next day it was loaded on a convoy, completely filling four heavy trucks. Two TICOM members, including First Lieutenant Selmer Norland, who would also go on to a long career at NSA, accompanied the equipment and soldiers back to England. There it was set up near Bletchley Park and quickly put into operation. It, or a working model, was later shipped back to Washington. The discovery of the Russian code breaking machine was a principal reason why both the U.S. and British governments still have an absolute ban on all details surrounding the TICOM operations.
All told, the TICOM teams salvaged approximately five tons of German Sigint documents. In addition, many cryptologic devices and machines were found and returned to Bletchley.
Equally important were the interrogations of the nearly 200 key German codebreakers, some of which were conducted at a secret location codenamed Dustbin. In addition to the discovery of the Russian Fish, another reason for the enormous secrecy surrounding TICOM may be the question of what happened to the hundreds of former Nazi code-breakers secretly brought to England. Were any of the war criminals given new identities and employed by the British or American government to work on Russian code breaking problems? Among those clandestinely brought into the United States was the top codebreaker Dr. Erich Huettenhain. "It is almost certain that no major cryptanalytic successes were achieved without his knowledge," said one TICOM document.
Among the surprises to come out of the interrogations was the fact that the Germans knew all along that Enigma was not totally secure. "We found that the Germans were well aware of the way the Enigma could be broken," recalled Howard Campaigne. "But they had concluded that it would take a whole building full of equipment to do it. And that's what we had. A building full of equipment. Which they hadn't pictured as really feasible."
In Washington, the TICOM materials were of enormous help in determining just how secure, or insecure, America's own cryptographic systems were. The picture painted by the documents and interrogations showed that while a number of lower-level systems had been read by German codebreakers, the most important ciphers remained impenetrable. "European cryptanalysts were unable to read any U.S. Army or Navy high-level cryptographic systems," the highly secret report said.
The Germans were never able to touch America's "Fish," a machine known as the SIGABA. Like the Fish, SIGABA was used for the Army and Navy's most sensitive communications. In fact, because TICOM showed that the SIGABA survived the war untouched by enemy codebreakers, it remained in service for some time afterward. It was finally taken out of service only because it did not meet the speed requirements of modern communications.
The TICOM report also indicated that other systems were not secure. One Army system and one Navy system were read for a short time. Both of the un-enciphered War Department telegraph codes were read by the Germans, and Hungary received photostats of War Department Confidential Code Number 2, probably from the Bulgarians. Also, thanks to a spy, Military Intelligence Code Number 11, which was used by the military attaché in Cairo, was read throughout the summer of 1942.
The most serious break was the solving of the Combined Naval Cypher Number 3, used by U.S. and Royal Navy convoy operations in the Atlantic; this Axis success led to many deaths. Other systems were also broken, but they were of less importance than the Allied breaks of Enigma and Fish.
By far the greatest value of TICOM, however, was not in looking back but in looking forward. With the end of the war, targets began shifting, the signals intelligence agencies dramatically downsized, and money became short. But at the start of the Cold War, as a result of TICOM, America had a significant lead. Not only did the U.S. code-breakers now have a secret skeleton key to Russia's Fish machine, it had a trapdoor into scores of code and cipher systems in dozens of countries. As a result of the German material and help from the British, for example, diplomatic communications to and from Afghanistan became "practically 100% readable." Thus, when Soviet officials discussed Asian diplomatic issues with the Afghan prime minister, the U.S. could listen in.
It was a remarkable accomplishment. At the outbreak of the war in Europe in 1939, the United States was attacking the systems of only Japan, Germany, Italy, and Mexico. But by the day the war ended, according to the TICOM report, "cryptanalytic attack had been directed against the cryptographic systems of every government that uses them except only our two allies, the British and the Soviet Union." Now readable, either fully or partially, were the encryption systems of Argentina, Belgium, Brazil, Bulgaria, Chile, China, Colombia, the Dominican Republic, Egypt, Ecuador, Ethiopia, Finland, France, Greece, Hungary, Iran, Iraq, Ireland, Italy, Japan, Lebanon, Mexico, the Netherlands, Peru, Portugal, Saudi Arabia, Spain, Switzerland, Syria, Thailand, Transjordan, Turkey, Uruguay, Venezuela, and Yugoslavia.
Between the attack on Pearl Harbor and August 1945, the Army's Signal Security Agency's Language Branch scanned more than 1 million decrypted messages and, of those, forwarded approximately 415,000 translations. But then it was over. Brigadier General W. Preston Corderman, chief of the Army codebreakers, was sure there would no longer be a need for much of a cryptanalytic effort. He therefore assembled the staff beneath the tall maple trees that gave his headquarters shade in the summer. The war was over, he told them, and so was their country's need for their services.
"Overnight, the targets that occupied most of the wartime cryptologic resources—Germany and Japan—had become cryptologic nonentities," said one NSA report. "One by one the radio receivers that had been faithfully tuned to enemy signals were switched off. Antenna fields were dismantled, equipment mothballed as station after station around the world ceased monitoring the airwaves, turned off the lights and padlocked the doors. Gone were the Army intercept stations at Miami, Florida; at New Delhi, India; at OSS Operations in Bellmore, New York; at Tarzana, California; and at Accra on the African Gold Coast. Silent were the Radio Intelligence Companies supporting General MacArthur in the Southwest Pacific and the Signal Service Companies in Europe."
The relative handful of American codebreakers who stayed on quickly shifted gears. The Soviet Union instantly became their number one target.
One key listening post not shut down was Vint Hill Farms Station. Known as Monitoring Station Number 1, it was located in the rural Virginia town of Warrenton. During the war, Vint Hill played a pivotal role in eavesdropping on enemy communications for thousands of miles in all directions. At war's end, 2,600 people stayed on, many of them intercept operators, to handle the transition from hot war to cold war.
They were able to eavesdrop on key Russian diplomatic and military communications sent over the Fish machine. "They intercepted printers at Vint Hill, Russian printers," said Colonel Russell H. Horton, who commanded the station shortly after the end of the war. "They had these ... circuits that had nine channels if I'm not mistaken. They had machines all hooked up so that they separated the channels and did all of the interception in Cyrillic characters." Horton added, "As far as I know, there was no effort against the Russians until after the war."
Although the fact was known to only a few, a small group of codebreakers had in fact been working on Russian code problems during the war. In 1943, American intelligence began to worry about a possible alliance between Nazi Germany and Russia as part of a comprehensive peace deal. Such a merger would have been a nightmare for the Allies. As a result, a few Army cryptanalysts were pulled away from work on German systems and assigned to a highly secret new unit with the goal of attempting to solve the enormously complex Soviet codes and ciphers.
Since 1939, thousands of encrypted Soviet messages, sent between Moscow and Washington, had been acquired from Western Union and other commercial telegraph companies. A major break occurred when it was discovered that identical code groups turned up in seven pairs of messages. To find even a single pair was a billion-to-one shot. Army codebreakers had discovered a "bust," an error or anomaly that opens a crack into the cipher system. Such a bust might be caused, for example, by a malfunction in a random-number generator. This bust, however, was caused by the Soviets reusing pages from one-time pads—the violation of a cardinal cryptographic rule. One-time pads had become two-time pads. Cecil Phillips, a former senior NSA official, played a key role in the early Soviet-watching program. "For a few months in early 1942," he said, "a time of great strain on the Soviet regime, the KGB's cryptographic center in the Soviet Union for some unknown reason printed duplicate copies of the 'key' on more than 35,000 pages . . . and then assembled and bound these one-time pads. . . . Thus, two sets of the ostensibly unique one-time pad page sets were manufactured."
The decision by the Soviet codemakers to duplicate the pages was likely the result of a sudden shortage of one-time pads, a result of Hitler's invasion of Russia in June 1941. To quickly fill the enormous demand for the pads, Russian cryptographers likely chose the easiest course: carbon paper. Suddenly production was doubled while, it was reasoned, security was diminished only slightly.
Phillips estimated that between 1942 and 1948, when the last onetime pad was used, more than 1.5 million messages were transmitted to Soviet trade and diplomatic posts around the world. Of those, American codebreakers obtained about a million, 30,000 of which had been enciphered with the duplicate pages. But despite the bust, days and weeks of frustrating work were required to squeeze out a clear-text message from a cipher text. Even then, usually the most they would have was a long, out-of-date message concerning such things as shipping schedules of the Soviet Purchasing Commission.
For more than thirty years the codebreakers worked on those messages. By the time the file drawer was closed for the last time, in 1980, they had managed to read portions of more than 2,900 Soviet diplomatic telegrams sent between 1940 and 1948. Codenamed Venona, the program was one of the most successful in NSA's history. It played a major role in breaking up key Soviet espionage networks in the United States during the postwar period, including networks aimed at the secrets of the atomic bomb.
On April 25, 1945, as TICOM officers began sloshing through the cold mud of Europe, attempting to reconstruct the past, another group of codebreakers was focused on a glittering party half the earth away, attempting to alter the future.
Long black limousines, like packs of panthers, raced up and down the steep San Francisco hills from one event to another. Flower trucks unloaded roses by the bushel. Flashbulbs exploded and champagne flowed like water under the Golden Gate. The event had all the sparkle and excitement of a Broadway show, as well it should have. The man producing it was the noted New York designer Jo Mielziner, responsible for some of the grandest theatrical musicals on the Great White Way. "Welcome United Nations," proclaimed the bright neon marquee of a downtown cinema. The scene was more suited to a Hollywood movie premiere than a solemn diplomatic event. Crowds of sightseers pushed against police lines, hoping for a brief glimpse of someone famous, as delegates from more than fifty countries crowded into the San Francisco Opera House to negotiate a framework for a new world order.
But the American delegates had a secret weapon. Like cheats at a poker game, they were peeking at their opponents' hands. Roosevelt fought hard for the United States to host the opening session; it seemed a magnanimous gesture to most of the delegates. But the real reason was to better enable the United States to eavesdrop on its guests.
Coded messages between the foreign delegations and their distant capitals passed through U.S. telegraph lines in San Francisco. With wartime censorship laws still in effect, Western Union and the other commercial telegraph companies were required to pass on both coded and uncoded telegrams to U.S. Army codebreakers.
Once the signals were captured, a specially designed time-delay device activated to allow recorders to be switched on. Devices were also developed to divert a single signal to several receivers. The intercepts were then forwarded to Arlington Hall, headquarters of the Army codebreakers, over forty-six special secure teletype lines. By the summer of 1945 the average number of daily messages had grown to 289,802, from only 46,865 in February 1943. The same soldiers who only a few weeks earlier had been deciphering German battle plans were now unraveling the codes and ciphers wound tightly around Argentine negotiating points.
During the San Francisco Conference, for example, American codebreakers were reading messages sent to and from the French delegation, which was using the Hagelin M-209, a complex six-wheel cipher machine broken by the Army Security Agency during the war. The decrypts revealed how desperate France had become to maintain its image as a major world power after the war. On April 29, for example, Fouques Duparc, the secretary general of the French delegation, complained in an encrypted note to General Charles de Gaulle in Paris that France was not chosen to be one of the "inviting powers" to the conference. "Our inclusion among the sponsoring powers," he wrote, "would have signified, in the eyes of all, our return to our traditional place in the world."
In charge of the San Francisco eavesdropping and codebreaking operation was Lieutenant Colonel Frank B. Rowlett, the protégé of William F. Friedman. Rowlett was relieved when the conference finally ended, and he considered it a great success. "Pressure of work due to the San Francisco Conference has at last abated," he wrote, "and the 24- hour day has been shortened. The feeling in the Branch is that the success of the Conference may owe a great deal to its contribution."
The San Francisco Conference served as an important demonstration of the usefulness of peacetime signals intelligence. Impressive was not just the volume of messages intercepted but also the wide range of countries whose secrets could be read. Messages from Colombia provided details on quiet disagreements between Russia and its satellite nations as well as on "Russia's prejudice toward the Latin American countries." Spanish decrypts indicated that their diplomats in San Francisco were warned to oppose a number of Russian moves: "Red maneuver . . . must be stopped at once," said one. A Czechoslovakian message indicated that nation's opposition to the admission of Argentina to the UN.
From the very moment of its birth, the United Nations was a microcosm of East-West spying. Just as with the founding conference, the United States pushed hard to locate the organization on American soil, largely to accommodate the eavesdroppers and codebreakers of NSA and its predecessors. The Russians, on the other hand, were also happy to have the UN on American soil—it gave them a reason to ship dozens of additional spies across U.S. borders.
Since the discovery of the Russian Fish machine by TICOM at the end of the war, and the ability to read a variety of diplomatic, KGB, and trade messages as a result of the Venona breakthrough on Soviet one time pads, American codebreakers had been astonishingly lucky. Virtually overnight they were placed in what NSA has called "a situation that compared favorably to the successes of World War II." For several years, American codebreakers were able to read encrypted Soviet armed forces, police, and industry communications and the agency could put together "a remarkably complete picture of the Soviet national security posture." But then, almost overnight in 1948, everything went silent. "In rapid succession, every one of these cipher systems went dark," said a recent NSA report, which called it "perhaps the most significant intelligence loss in U.S. history." It forever became known at NSA as Black Friday.
For American codebreakers, the lights could not have gone out at a worse time. In late June 1950, North Korean forces poured across the 38th Parallel into the south, launching the Korean War. Once again, as with Pearl Harbor, America was caught by surprise.
A year before the attack, the Army, Navy, and Air Force code-breaking organizations had been combined into a single unit, AFSA. But instead of establishing a strong, centralized organization to manage the growing worldwide signals intelligence operations, each service was allowed to retain control of both intercept and code breaking activities. That left little for the director of AFSA to direct. Nor could he even issue assignments to field units. They would first have to pass through each of the services, which could then accept them, change them, or simply ignore them. Herbert L. Conley, who was in charge of Russian traffic analysis at AFSA in the late forties, and later headed up Russian code-breaking at NSA, likened the organization to a "three-headed monster." "He couldn't control anything outside of the buildings that were occupied," he said of the director.
In the weeks leading up to the attack, Korea barely registered as a Sigint target for AFSA. Out of two priority lists, North Korea was number fifteen on the secondary list. From listening posts at Kamiseya, Japan, and several other locations, most of the intercept activity was directed at Russia. Communist China was also a high priority, with eighty-seven intercept operators and analysts focused on it. But because AFSA had not broken any important Chinese cipher systems, most personnel concentrated on traffic analysis, the examination of the message's "external indicators," such as its date and "to" and "from" lines. North Korea, on the other hand, was targeted by just two intercept operators at the time the war broke out. In all, they had collected a paltry two hundred messages, and none of those had been processed. "AFSA had no Korean linguists, no Korean dictionaries, no traffic analytic aids, and no Korean typewriters," said a later NSA analysis.
Despite the limited resources, clues were there. Buried in stacks of intercepted Soviet traffic as far back as February were messages pointing to large shipments of medical supplies going from Russia to Korea. Other messages, about the same time, revealed a sudden and dramatic switch toward targets in South Korea by Soviet radio direction-finding units.
Suddenly, at 3:30 on the morning of June 25, 1950, Joseph Darrigo, a U.S. Army captain and the only American on the 38th Parallel, was jarred awake by the teeth-rattling roar of artillery fire. At that moment North Korean ground forces, led by 150 Soviet T-34 tanks, began their massive push into South Korea. Darrigo managed to escape just ahead of the advancing troops and spread the alarm. "AFSA (along with everyone else) was looking the other way when the war started," said a recent, highly secret NSA review. The first word to reach Washington came from a news account by a reporter in Seoul.
Within days, the North Korean Army had captured Seoul and continued to steamroll south, seeking to unify the peninsula under the flag of communism. In response, American troops were quickly dispatched to provide assistance to South Korea as part of a United Nations force. By the end of the first week, 40,000 South Korean soldiers had been killed, captured, or declared missing in action.
Following the attack, AFSA began a quick push to beef up its ranks. The number of intercept positions targeting North Korean traffic jumped from two to twelve. Any signals even remotely North Korean were transmitted back to AFSA headquarters in Washington, arriving ten to twelve hours after intercept. Soon, new messages were arriving hourly and lights were burning around the clock.
Nevertheless, cryptanalysis was virtually nonexistent. In fact, the first few decrypts of enciphered North Korean air traffic were produced not by professional codebreakers but by an uncleared U.S. Army chaplain using captured codebooks. Seconded into Sigint duty, Father Harold Henry had spent a number of years in Korea, where he learned the language. Most analysts instead concentrated on traffic analysis and plaintext intercepts—highly useful because of poor communications security by the North Koreans during the early part of the war. Among the messages sent in the clear were secret battle plans.
Adding to the problems, it was three months before a small advanced Sigint unit actually arrived on the Korean peninsula. Radio direction finding was greatly hampered by the mountainous terrain. Then there were the supply shortages, outmoded gear, difficulties in determining good intercept sites, equipment ill-suited to frequent movement over rough terrain, and a significant lack of translators.
From the beginning, the ground war went badly. By the end of July, the Eighth Army, led by General Walton H. Walker, had been forced into a boxlike area known as the Pusan Perimeter, so named because it surrounded the southeastern port of Pusan. "When we got into the . . . Perimeter, you never saw a more beat-up bunch of soldiers," recalled former PFC Leonard Korgie. "The North Koreans had hellish numbers and equipment. We were very, very thin in both."
Walker's one advantage was a constant supply of Sigint, which provided him with such vital information as the exact locations of North Korean positions. Armed with this intelligence, he was able to maximize his limited men and resources by constantly moving them to where new attacks were planned. Finally, following MacArthur's daring amphibious landing at Inchon, a port located behind enemy lines, Walker's men broke out of their box and joined in the attack, putting North Korea on the defensive.
In one sense, Sigint in Korea was like a scene from Back to the Future. After planting a number of sound-detecting devices forward of their bunkers to give warning of approaching troops, ASA soldiers discovered that the devices also picked up telephone calls. So they began using them for intercept—a practice common during World War I but long forgotten. This "ground-return intercept," using the principle of induction, enabled the ASA to collect some Chinese and Korean telephone traffic. The downside, however, was that in order to pick up the signals the intercept operator had to get much closer to enemy lines than normal, sometimes as close as thirty-five yards.
"One of our problems in Korea was linguists, there were so few," said Paul Odonovich, an NSA official who served in Korea with the Army 26 Security Agency. Odonovich commanded a company of intercept operators on the front lines. Sitting in antenna-bedecked vans, they would mostly eavesdrop on North Korean "voice Morse," an unusual procedure whereby the North Korean military would read the Morse code over the communications channels rather than tap it out with a key. "They used the singsong 'dit-dot-dit-dit' business," said Odonovich.
Other units conducting low-level voice intercept (LLVI), as it was known, operated out of jeeps and bunkers close to the front lines. The intelligence was then disseminated directly to combat units. By the end of the war, twenty-two LLVI teams were in operation. Air Force intercept operators also had some successes. Operating from small islands off North Korea, Sigint units were able to intercept North Korean, Chinese, and Soviet instructions to their pilots. The intercept operators would then disguise the intelligence as "radar plots" and pass them on in near real time to U.S. pilots operating over North Korean territory. Once they received the information, their "kill ratio" increased significantly.
After the battle began, the most important question was whether China would intervene. Since the end of World War II, Army Sigint specialists had engaged in a haphazard attack on Chinese communications. In 1945, General George Marshall attempted to bring Nationalist leader Chiang Kai-shek and Communist boss Mao Tse-tung to the negotiating table. At Marshall's request, a small group of intercept operators eavesdropped on both sides during the talks.
But the operation was less than a success. A team set up in Nanjing to intercept Nationalist communications was hampered by unreliable electrical power. Another, which targeted Communist links from a listening post in Seoul, was plagued with "poor hearability." Ironically, as the United States struggled, the British had been secretly listening to Chinese Communist communications for years. From 1943 until 1947, the Government Code and Cypher School successfully monitored a link between Moscow and Mao's headquarters in Yan'an, China. But because the link was part of a clandestine Soviet network, the decision was made to keep the Americans in the dark until March 1946.
Nevertheless, from the messages that the United States was able to intercept, it was clear that the two groups preferred to settle their differences on the battlefield rather than at the conference table. As a result, the Marshall mission was withdrawn in 1946. Thereafter, ASA dropped its study of Chinese Communist military ciphers and communications and turned its attention almost exclusively toward Russia. It would prove a serious mistake. Three years later, in 1949, Mao triumphed and Chiang fled to the island of Formosa.
About the same time, a small team of Chinese linguists led by Milton Zaslow began eavesdropping on and analyzing Chinese civilian communications—private telephone calls and telegrams. Unencrypted government messages would also travel over these lines. Beginning in early summer 1950, AFSA began developing "clear and convincing evidence" that Chinese troops were massing north of the Yalu River.
In May and June, Sigint reports noted that some 70,000 Chinese troops were moving down the Yangtze River in ships toward the city of Wuhan. The next month a message intercepted from Shanghai indicated that General Lin Piao, the commander of Chinese army forces, would intervene in Korea. Later reports noted that rail hubs in central China were jammed with soldiers on their way to Manchuria. By September, AFSA had identified six field armies in Manchuria, near the Korean border, and ferries on the Yalu River were being reserved for military use.
All of these reports were fully available to the Joint Chiefs of Staff, the White House, and to General Douglas MacArthur, the commander of the UN forces. Nevertheless, when asked by President Truman on October 15 about the chances of Chinese intervention, MacArthur replied, "Very little."
The indications continued. On October 21, AFSA issued a Sigint report stating that twenty troop trains were heading toward Manchuria from Shanghai. Then, on November 7, AFSA intercepted a radiotelephone call made by an East European in Beijing. He reported that orders had been issued allowing every Chinese soldier to volunteer to fight in Korea, saying, "We are already at war here." That same month, intercept operators picked up an unencrypted order for 30,000 maps of Korea to be sent from Shanghai to the forces in Manchuria.
Finally, intercepts during the first three weeks of November revealed that Beijing was in a state of emergency, with authorities sponsoring mass demonstrations demanding intervention, imposing more stringent censorship, improving air defense, and commanding that any soldier or officer could volunteer to serve in Korea. A medical headquarters urgently ordered troops in Manchuria to receive immunizations for diseases that were prevalent in North Korea—smallpox, cholera, and typhoid fever. AFSA reports demonstrated clearly that the Chinese were making extensive preparations for war.
But despite the many Sigint clues, U.S. and South Korean forces were once again caught by surprise. Early on the bitter-cold morning of November 26, with trumpets braying, thirty Chinese divisions surged across the North Korean border and forced U.S. and South Korean armies to make a precipitous retreat southward, costing the lives of many American soldiers.
"No one who received Comint product, including MacArthur's own G-2 [intelligence chief] in Tokyo, should have been surprised by the PRC intervention in the Korean War," said a recent, highly classified NSA review. The review then pointed a finger of blame for the disaster directly at MacArthur. "During the Second World War, MacArthur had disregarded Comint that contradicted his plans," it said. "MacArthur's zeal [to press ahead] to the Yalu probably caused him to minimize the Comint indicators of massive PRC intervention just as he had earlier minimized 'inconvenient' Comint reports about the Japanese. He thus drove his command to great defeat in Korea."
By mid-1951, with the 38th Parallel roughly dividing the two sides, ASA headquarters was established in the western suburbs of Seoul, on the campus of Ewha College, the largest women's school in Asia. There, traffic analysts put together a nearly complete Chinese army order of battle. Also, when truce negotiations began in July 1951, ASA units eavesdropped on meetings among the North Korean negotiating team. But that same month, the earphones of most of the intercept operators went silent as the North Koreans switched much of their radio communications to the security of landlines. NSA later attributed this caution to secrets allegedly passed to the Russians by former AFSA employee William Weisband.
Toward the end of the war, there were some tactical successes. By 1952, AFSA had broken a number of Chinese cipher systems. "The . . . last three major pushes that the Chinese had against us, we got those lock, stock, and barrel, cold," recalled Odonovich. "So that when the Chinese made their advances on our positions they were dead ducks . . . we had the code broken and everything."
But critical high-level communication between and among the Chinese and North Koreans was beyond the AFSA codebreakers' reach. Gone was the well-oiled machine that had helped win World War II. In its place was a confusing assortment of special-interest groups, each looking upon the other as the enemy; no one had the power to bring them together. "It has become apparent," complained General James Van Fleet, commander of the U.S. Eighth Army in June 1952, "that during the between-wars interim we have lost, through neglect, disinterest and possibly jealousy, much of the effectiveness in intelligence work that we acquired so painfully in World War II. Today, our intelligence operations in Korea have not yet approached the standards that we reached in the final year of the last war." A year later NSA director Ralph Canine, an Army lieutenant general, concurred with Van Fleet's observation.
So bad was the situation that in December 1951 the director of the CIA, Walter Bedell Smith, brought the problem to the attention of the National Security Council. In his memorandum, Smith warned that he was "gravely concerned as to the security and effectiveness with which the Communications Intelligence activities of the Government are being conducted." He complained that American Sigint had become "ineffective," as a result of the "system of divided authorities and multiple responsibilities."
Smith then discreetly referred to the mammoth security breach, blamed on Weisband, that had led the Soviets to change their systems. "In recent years," he said, "a number of losses have occurred which it is difficult to attribute to coincidence." To preserve what he called "this invaluable intelligence source"—Sigint—Smith called on Truman to ask Secretary of Defense Robert A. Lovett and Secretary of State Dean G. Acheson to conduct a "thorough investigation" of the agency. Three days later, on December 13, 1951, Truman ordered the investigation.
Appointed to head the probe was George Abbott Brownell, a fiftythree-year-old New York attorney and former special assistant to the secretary of the Air Force. Over six months, Brownell and his committee of distinguished citizens took AFSA apart and put it together again. In the end, they viewed AFSA as a "step backward." By June 13, 1952, when he turned his report over to Lovett and Acheson, Brownell had a blueprint for a strong, centralized new agency with a director more akin to a czar than to the wrestling referee the post resembled. Both secretaries approved and welcomed the independent review and set about carrying out its recommendations.
Four months later on October 24, Lovett, David K. Bruce from the State Department, and Everett Gleason of the NSC entered the Oval Office for a 3:30 off-the-record meeting with the president. There, Truman issued a highly secret order scrapping AFSA and creating in its place a new agency to be largely hidden from Congress, the public, and the world. Early on the morning of November 4, as Truman was leaving a voting booth in Independence, Missouri, the National Security Agency came to life. But few gave the new agency much hope. "The 'smart money' was betting that the new organization would not last much longer than AFSA," scoffed one official.
That night, Dwight David Eisenhower was elected the thirty-fourth president of the United States.
next.... NERVES
No comments:
Post a Comment