<<

. 53
( 118 .)



>>

look into the activities of the federal government, including the
intelligence community, came away stunned. "Monetary considerations
should be waived," they recommended to Eisenhower, "and an effort at
least equal to the Manhattan Project [which built the atomic bomb
during World War II] should be exerted at once" to produce high-level
signals intelligence. The Pentagon authorized NSA "to bring the best
possible analytical brains from outside NSA to bear on the problem (if
they can be found)." The President's Board of Consultants on Foreign
Intelligence Activities called NSA "potentially our best source of accurate
intelligence." Finally, the White House's Office of Defense Mobilization
recommended "that the Director of the National Security Agency be made
a member or at least an observer on the Intelligence Advisory
Committee."
Soon NSA went from lean to fat. Its funding rose above $500 million,
more than half the entire national intelligence budget. The exploding
costs greatly concerned even Eisenhower himself. "Because of our having
been caught by surprise in World War II," he said, "we are perhaps
tending to go overboard in our intelligence effort." During a meeting of
the Special Comint Committee in the Oval Office, Treasury Secretary
George Humphrey, an old quail-shooting friend of Eisenhower's,
exclaimed that he "was numb at the rate at which the [NSA] expenditures
were increasing." But with regard to NSA, Eisenhower made an exception
to his financial anxiety. "It would be extremely valuable if we could break
the Soviet codes," he said.
Also at the meeting was fifty-four-year-old James R. Killian, Jr. As
chairman of the President's Board of Consultants on Foreign Intelligence
Activities and president of the Massachusetts Institute of Technology, the
Eisenhower adviser was intimately familiar with the need for good
intelligence. A few years earlier he had conducted a highly secret study


299
for Eisenhower on the risks posed to the nation by a surprise attack.
Now, in its formal report to the president, the board called for an even
greater effort against Russian encryption systems. "In our judgment the
intelligence 'breakthrough' which would yield us greatest dividends
would be the achievement of a capability to break the Soviet high-grade
ciphers," it said.
Killian offered a suggestion. "An essential step in seeking a solution to
this problem," he urged, "would be a successful mobilization of the best
available talent in the country to search out the most promising lines of
research and development." Eisenhower approved the recommendation,
and Dr. William O. Baker, vice president for research at Bell Labs, was
appointed to head the scientific study into ways to improve NSA's attack
on Soviet high-grade ciphers. On February 10, 1958, the final Baker
Report was hand-delivered to Eisenhower. Baker reported his
committee's view that NSA "was providing the best intelligence in the
community." NSA's intercept capability and its analysis of electronic and
telemetry intelligence greatly impressed the committee, and Baker
recommended that NSA have complete dominance over all electronic
intelligence (Elint). Thus his report settled a long battle between NSA and
the Air Force for control of the rapidly growing field. But the Baker
Committee also believed that foreign codemakers had outpaced NSA's
codebreakers and expressed its skepticism of NSA's abilities in
cryptanalysis.
Killian also pushed Eisenhower to place great emphasis on the
development by NSA "of machines and techniques for speeding up the
sifting out of important items from the great mass of information that is
accumulated daily from Communications Intelligence sources." This also
Eisenhower carried out.
Among the key areas the Baker Committee suggested concentrating
on was Soviet ciphony, or scrambled voice communications. Two decades
earlier, in 1939, Franklin D. Roosevelt and Winston Churchill carried on,
over a scrambler phone, a series of highly sensitive discussions regarding
the growing war in Europe. At the White House, the telephone link was
in the basement, and in London it was in Churchill's underground war
cabinet rooms.
The system had been developed by Bell Telephone. Known as the A-3,
it worked by breaking up the frequency bands and scattering the voice
impulses at one end and then reconstructing them, like pieces of a jigsaw
puzzle, at the other end. Roosevelt's voice first traveled to an AT&T
security room in New York. There the signal was mangled into gibberish
before being transmitted to England on an undersea cable. In London, it
was electronically stitched back together.
Barely had Roosevelt received his first call on the machine when



300
Germany's post minister, who had overseen the tapping of the undersea
cable from England to the United States, began looking for ways to break
into the system. Working without blueprints or any idea what the actual
system looked like, the engineer nevertheless succeeded in "breaking" the
cipher system within only a few months. Thereafter, Hitler was receiving
transcripts on his desk of some of the most secret conversations of the
war. Among the results was a disastrous prolongation of the war in Italy.
During the 1960s, NSA's inability to break high-level Soviet codes was
becoming its biggest secret. CIA director John McCone became so
concerned that in 1964 he asked Richard Bissell to look into the
problem. Bissell was one of the CIA's keenest scientific minds, one of the
key people behind the U-2, the SR-71, and early reconnaissance
satellites. Unfortunately, because of his involvement in the Bay of Pigs
debacle he was fired by President Kennedy. Bissell then went to the
Institute for Defense Analysis, which had long run NSA's secret think
tank, the IDA Communications Research Division. After Bissell left IDA,
about 1964, McCone asked him to conduct a special study of NSA's most
sensitive codebreaking efforts against high-level Soviet cipher systems.
The idea of the CIA sending an outsider to poke into NSA's deepest
secrets horrified many at the codebreaking agency.
"I finally did produce a report which went to the DCI [Director of
Central Intelligence] and NSA," said Bissell, "though it was so secret I
couldn't even keep a copy of it under any circumstances and I don't
know whether I was even allowed to read it again. But they [NSA] went
around and told the DCI, who had commissioned it and to whom it was
addressed, that he had to turn his copy in to the NSA, which he refused
to do." A later CIA director would occasionally ask top NSA officials
whether they had made any breakthroughs, but the answer was usually
vague. "I could never tell how close they were to doing this with the
Russians," he said. "They would say they were close, but they never did it
as far as I was aware of."
In default of effective cryptanalysis, for the most part A Group
analysts relied on traditional traffic analysis, Elint, and unencrypted
communications for their reports. Another source of Soviet intelligence
came from breaking the cipher systems of Third World countries. Often
after meetings with Soviet officials, the Third World diplomats would
report back to their home countries over these less secure systems.
By the late 1970s the science of ciphony had progressed considerably,
but it was still considered far more vulnerable than encrypted written
communications. In NSA's A4 section, the Russian ciphony problem was
given the codename Rainfall. Day after day, codebreakers assigned to
Rainfall searched endlessly for a "bust," an error that would act as a
toehold in their climb up the cryptanalytic mountain. At last, in the late
1970s, they began to find it. "When they went bust," said one of those


301
involved in the project, "the Soviet encryption failed so they couldn't set
up the encryption. In an attempt to reestablish the encrypted link, they
had to go plaintext. This became a major thing. People would run into
where we were working and you'd get around nine or ten people hovering
around a receiver. It was a major event to hear in clear text what
normally would have been encrypted. This was real time."
When one or both ends of a scrambled conversation failed to
synchronize correctly, the encryption would fail. In that case the
Russians would have to try to fix the problem before going ahead with
their conversation. But occasionally, either because they did not realize
the encryption had not kicked in or simply out of laziness, the
transmission would continue in the clear. At other times the parties
would begin discussing the problem and in so doing give away important
secrets of the system, such as keying information. As time went on, the
Rainfall cryptologists discovered enough toeholds in the Soviet scrambler
phone so that they were able to break the system even when it was
properly scrambled.
Another problem was how to intercept the scrambler-phone signal
and other Soviet communications without the Russians knowing. In
trying to solve this problem, for twenty years NSA had been moving more
and more toward space-borne eavesdropping. The process had begun on
the back of a placemat in a Howard Johnson's restaurant during a
snowstorm.


"One good intercept is worth $5 million," Robert O. Aide of NSA's
Research and Development Group (RADE) told his colleague Nate Gerson
in the late 1950s. More than four decades later, as a senior cryptologic
scientist at NSA, Gerson recalled that the urgency of obtaining Sigint on
Soviet space activities heightened greatly after the successful Russian
launch of Sputnik 1 in 1957. Of key concern was telemetry, the revealing
signals transmitted from the missile to the launch center. "Aide kept
firing me up," said Gerson, "about the value to NSA of receiving the
telemetry."
Other people were exploring the same problem in unconventional
ways. At a meeting with Eisenhower in 1959, Killian suggested placing
eavesdropping balloons at six points around the earth, at an altitude of
about fourteen miles. "This has great promise for monitoring Soviet
missile firings," he said. The reason was that "sound ducts" occur at that
altitude. "At this level," Killian said, "sound tends to stay in the layer of
air." Eisenhower thought the idea "splendid." However, he was worried
that the secret might get out; he commented on the way "irresponsible
officials and demagogues are leaking security information."
To Gerson, the problem was capturing the missile's signal. Because


302
the signal was line-of-sight and the launch pad was far inland, it was
difficult to intercept with peripheral ferret flights. Gerson explored ways
to create atmospheric conditions that, like a mirror, would reflect the
signal long distances. Once the signal had been reflected beyond Soviet
borders, land-based or airborne collectors could intercept it. In 1959
Gerson submitted his report, "Six Point Program for Improved Intercept,"
was given an initial $1 million in research money, and began to
experiment.
An intercept station was set up in the Bahamas. Its target was an
unsuspecting television station in Shreveport, Louisiana, about 1,500
miles away. (Television broadcast signals are line-of-sight.) At a certain
point over the southwestern United States, a rocket that had been
launched from Eglin Air Force Base in Florida detonated into the
atmosphere a chemical bomb containing aluminum oxide and cesium
nitrate. Cesium nitrate is hazardous. Users are warned, "Do not breathe
dust, vapor, mist, or gas; do not get in eyes, on skin or clothing, and
obtain medical attention if it is inhaled." Nevertheless, no one thought to
warn residents under the bomb.
As the toxic cloud drifted over Shreveport, the television signals
bounced off the heavy particles and were intercepted at the NSA listening
post in the Bahamas. "The experiments were successful and ultimately
allowed reception of TV signals far beyond the line-of-sight," said Gerson.
"The TV signals had been reflected from the electron cloud produced by
ionization of the chemical mixture. Reception persisted for about sixty
minutes."
Continuing with his experiments, Gerson next toyed with the idea of
launching a large reflector into space, off which the Soviet telemetry
signals would bounce down to a listening post. Then Gerson and an NSA
colleague "extended the calculations to include reflections from [that is,
signals bouncing off] the moon," he said, "and as an afterthought, from
Mars and Venus. We were both somewhat surprised with the results; the
concept was feasible if a sufficiently high-gain antenna were available."
Later, in the early 1960s, the Pentagon's Advanced Research Projects
Agency (ARPA) began funding construction of the mammoth Arecibo
Ionosphere Observatory in Puerto Rico. A scientific antenna used to
explore the earth's ionosphere and surrounding space, it was built over a
large sinkhole, which acted as a perfect base for the antenna's 900-foot-
plus dish. The dish's size ensured enormous receiving capability.
However, because it used a natural sinkhole, the antenna itself was fixed
in place; only the 900-ton feed platform that was suspended above the
bowl-shaped reflector could move.
Gerson thought the Arecibo dish would be a perfect antenna to
capture Soviet signals as they drifted into space, bounced off the moon,



303
and were reflected back to earth. He approached the director of ARPA,
Charles Herzfeld, to broach the possibility of allowing NSA to experiment
with the antenna. "Herzfeld told us in no uncertain terms that AIO
[Arecibo Ionosphere Observatory] had been funded as a wholly scientific
and open facility," said Gerson, "and would not be allowed to undertake
classified studies, and that it was presumptuous of us to ask." But
Herzfeld later gave in, and NSA began using the antenna under the cover
of conducting a study of lunar temperatures.
(Indeed, ARPA suddenly became extremely helpful to NSA, even to the
point of offering to nuke the Seychelles Islands for them. At one point,
while NSA was planning its intercept operation at Arecibo, Gerson
mentioned that while the antenna was ideal, the location was bad. The
best place, he said, would be the Seychelles, in the Indian Ocean.
"[William H.] Godel of ARPA later approached me," recalled Gerson, "and
offered to construct a scooped antenna for NSA, in the Seychelles or
elsewhere. A nuclear detonation would be employed [to create a giant
hole for the antenna's dish] and ARPA guaranteed a minimum residual
radioactivity and the proper shape of the crater in which the antenna
subsequently would be placed. We never pursued this possibility. The
nuclear moratorium between the U.S. and the USSR was signed
somewhat later and this disappeared.")
NSA officials were amazed with the results at Arecibo. Just as
anticipated, the sensitive Russian signals drifted into space, ricocheted
off the moon, and landed, like a ball in the pocket of a pool table, in the
Arecibo dish on the other side of the planet. "After just one week of
operation," said Gerson, "we intercepted Soviet radar operating on the
Arctic coast." He added, "As a byproduct of my involvement, I could never
look at the moon again without thinking of our experiment."
About the same time, someone else at NSA developed equipment to
electronically trick Soviet satellites. Signals secretly transmitted to the
satellites would induce them to broadcast information down to where
NSA intercept operators could record it. The spoofing equipment was
placed at a field station, but Gerson and Donald H. Menzel, the director
of Harvard University's observatory, objected. Menzel was serving as an
NSA consultant. "We were both bothered about the precedent," said
Gerson. "It could prove self-defeating and result in constant electronic
tampering with the other's satellites. By the end of the summer 1960, the
equipment was disabled to prevent even an accidental occurrence of
tampering."
As Nate Gerson was looking for ways to snare elusive Soviet signals off
the moon, so was the Naval Research Laboratory. But rather than use
the limited Arecibo dish or nuke the Seychelles, the NRL was prowling
the fog-layered hollows of West Virginia. Finally, in a remote Allegheny
cranny of green washboard hills, they found the perfect place: Sugar


304
Grove, population forty-two. Nestled deep in the wooded and
mountainous South Fork Valley of Pendleton County, Sugar Grove was,
by law, quiet. Very quiet. To provide a radio-quiet zone for deep-space

<<

. 53
( 118 .)



>>