THE CODEBREAKERS

by DAVID KAHN

  None of these is a part of the American cryptologic organization, though all participate in its activities. That organization is not exhausted by N.S.A., its advisory board, and the agencies that feed it ideas and information. The creation of A.F.S.A. did not abolish the individual cryptologic agencies of the armed services. Though subject to N.S.A. in technical matters, they remain, as units of the Army, the Navy, and the Air Force, administratively separate from it.

  Oldest, and probably the most direct ancestor of N.S.A. in view of its contributions to N.S.A. of the core personnel of Friedman, Kullback, Sinkov, and Rowlett, is the Army Security Agency. It can trace some functions back to G.2 A.6 and the Code Compilation Section, both of the A.E.F., but administratively it stems, though tortuously, from Yardley’s MI-8 and his Black Chamber and from Friedman’s two-man cryptographic bureau in the War Department. As a well-defined unit, it began with the creation of the Signal Intelligence Service in 1929. It continued through the war, changing its name first to the Signal Security Service, then to the Signal Security Agency. On September 15, 1945, a few days after the war ended, the War Department detached the agency from the Signal Corps and placed it within the Intelligence Branch (which had tried at least four times during the war to steal it). It was renamed the Army Security Agency and was given authority over all Army cryptologic units, which had previously functioned independently under theater commanders and merely with the advice of the Signal Security Agency.

  In February, 1949, Army Regulation 10-125 set forth these as some of A.S.A.’s responsibilities:

  … the Chief, Army Security Agency, formulates and implements plans, policies, and doctrine on communication intelligence and communication security for the Army, and is specifically responsible for the following: …

  b. Production of communication intelligence for the Department of the Army.

  c. Investigation of the means employed for clandestine communications; and the preparation, detection, and processing of secret inks, microphotographs, and open codes and ciphers.

  d. Technical supervision of communication security activities of the Department of the Army, including cryptocenter activities, programs of cryptographic instruction, and surveillance of friendly radio and wire traffic….

  j. Preparation, production, storage, distribution, and accounting of all registered cryptomaterial, together with the publication of instructions necessary for the use, handling, and safeguarding of such material, except in such cases where these duties may otherwise be specifically assigned.

  Other responsibilities cover command of A.S.A. installations and units, liaison, preparation of publications, conduct of training programs, supervision of the Army Security Reserve, and advising the Department of the Army.

  The establishment of A.F.S.A. and N.S.A. must have given these duties a restricted meaning. Nevertheless, so greatly had communications, and consequently cryptology, expanded, that on April 14, 1964, the Army redesignated the agency as a major field command. The published description is as uninformative as that of N.S.A.: “The Commanding General, United States Army Security Agency, is responsible for the operations, training, administration, services, and supply of all units, personnel, activities, and installations under his command throughout the world. He performs specialized technical functions relating to the national security.”

  Two of his chief customers must be the Assistant Chief of Staff for Intelligence (G-2), in whose hands are placed staff responsibility for “communications intelligence, electronic intelligence, communications security, and electronic security …; Army cryptologic functions,” and the Chief of Communications-Electronics (formerly the Chief Signal Officer), who advises the Chief of Staff on “communications, including pertinent communications security.” The Chief of Communications-Electronics is assisted in his function of supplying cryptographic equipment to the Army by the United States Army Signal Communications Security Agency, which buys, distributes, registers, stores, and repairs that equipment. It follows each item from point to point with forms bearing titles like “Cryptomaterial Distribution Summary Record” and “Cryptomaterial Consolidated Flyleaf Receipt.” Headquarters of the U.S. Army Security Agency remains at the Arlington Hall Station in Virginia.

  The Navy’s cryptologic agency remains buried in the Office of Naval Communications, and little more is known of it beyond its name—Naval Security Group—and its location at Nebraska Avenue. Presumably it feeds information to the Office of Naval Intelligence, and, like the other agencies, its radio stations do intercept work for N.S.A. On December 31, 1963, the Navy had 10,701 men performing cryptologic duties, both in the Navy itself and detached to N.S.A., or about 1 man in 70.

  The United States Air Force Security Service, by contrast, actually issues a press kit. Activated in October, 1948, it is now a major command of the Air Force, headquartered in a spanking new three-story building in the shape of an inverted U at Kelly Air Force Base, San Antonio, Texas. It operates more than 50 units in 14 nations under four geographical subdivisions: European Security Region at Frankfurt, Pacific Security at Hawaii, the 6940th Security Wing at Goodfellow Air Force Base, Texas, and the 6981st Security Group at Elmendorf Air Force Base, Alaska. It also trains specialists at the U.S.A.F.S.S. Technical School at Goodfellow.

  Its press kit describes its operations in this way:

  Foreign nations who cherish harmful designs against the Free World are constantly seeking useful information regarding the U.S. Aerospace forces. A prime target for such information seekers is the United States Air Force communications system.

  The primary responsibility for insuring that these nations are denied access to information transmitted over U.S.A.F. communications facilities is vested in the United States Air Force Security Service (U.S.A.F.S.S.). This mission is one of communications/electronic surveillance—or more simply stated “providing communications security for the Air Force.”

  First of all, A.F.S.S. technicians attempt to develop and supply the necessary techniques and specialized equipment needed to safeguard classified information being transmitted by electrical means by the Air Force.

  Secondly, these technicians monitor and analyze unclassified Air Force electrical communications to determine the amount of information of intelligence value that can be derived from these communications.

  Finally, A.F.S.S. reports to the originator of such communications the information developed and any procedural discrepancies noted, and they make the necessary recommendations for securing these communications against exploitation by unauthorized agents or agencies.

  All this is performed, of course, under the general guidance of the N.S.A. Thus the codebooks and the lists of authenticators that U.S.A.F.S.S. produces must conform to N.S.A. policy. And the United States Government Organization Manual adds to a one-sentence statement of the U.S.A.F.S.S. monitoring duties this second sentence: “Additionally, U.S.A.F. Security Service units occasionally conduct research in communication phenomena in support of various elements of the U.S. Government.” That, of course, is a perfectly marvelous euphemism for interception. Some of the material must go to N.S.A., some to the Air Force Assistant Chief of Staff, Intelligence (A-2).

  The Army Security Agency, and probably the other two service cryptologic agencies as well, maintains reserve components as a pool of cryptologic talent in time of mobilization or emergency. During the Berlin crisis of 1961, three Army Security Agency units were among the first reservists to be called up. They were stationed at the A.S.A. School at Fort Devens, Massachusetts—though one battalion had to come all the way from California. The others were the 197th A.S.A. Company from New York and the 324th A.S.A. Battalion, with members from Chicago.

  The chiefs of the three military cryptologic agencies serve under the commanders of their own armed services, who also sit as the Joint Chiefs of Staff, which serves as the military staff of the Secretary of Defense. The staff of the Joint Chiefs includes a Directorate for Communications-Electronics (j-6), whose Security a
nd Electronic Warfare Division prepares cryptologic plans. These must be individual programs for specific operations, presumably laid out under principles established by N.S.A. The Director of Communications-Electronics is assisted in his work by the Military Communications-Electronics Board, comprised of the chief communications officers of the Army, Navy, Air Force, and Marine Corps. Among its 11 part-time panels of military personnel is one on security and cryptography and another on electronic warfare. Also within the purview of the Joint Chiefs is the Defense Intelligence Agency, which undoubtedly receives, through the service intelligence units, information that has originated in the service cryptologic agencies as well as in the National Security Agency.

  N.S.A., however, does not fall under the Joint Chiefs of Staff. Under an arrangement that became effective June 15, 1963, the N.S.A. director reports to that one of the Assistant Secretaries of Defense who serves as the Deputy Director of Defense Research and Engineering. He in turn reports to the Secretary of Defense, who sits on the National Security Council, which advises the President on domestic, foreign, and military policies relating to the national security. The other members of the National Security Council are the President, the Vice President, the Secretary of State, whose department is one of N.S. A.’s major customers, both for intelligence and for security policies, and the Director of the Office of Emergency Planning, which handles civil defense and civil mobilization. This office, which comes directly under the President, includes, as an assistant director, a Director of Telecommunications Management. As the President’s Special Assistant for Telecommunications, he coordinates the telecommunication activities of the government and thus may be involved with cryptologic activities in a very general way.

  Serving the National Security Council is the Central Intelligence Agency, which correlates information from the several branches of the intelligence community and presents it to the Council. There is a great interchange of this information among all members and at all levels of the intelligence community. Thus, N.S.A. may feed some intelligence to C.I.A. and some to State, and may in turn receive cribs to solution from the latter and some actual cipher keys from the former. Surprisingly, C.I.A. does some cryptanalysis itself (Rowlett worked for them for a while), and the F.B.I. Cryptanalytical and Translation Section attacks spy ciphers, such as Reino Hayhanen’s message found in the hollow nickel.

  Organizational chart of chief agencies of the United States involved in crypto logy as producers, consumers, or critics

  The interchange of information is controlled by the United States Intelligence Board, which advises C.I.A. and acts as a board of directors for the intelligence community. The director of the National Security Agency sits on this board. Other members include the director of the Defense Intelligence Agency, the heads of G-2, O.N.I., and A-2, the director of the State Department’s Bureau of Intelligence and Research, representatives of the F.B.I., and the Atomic Energy Commission, and the Director of Central Intelligence, who chairs the board.

  Watchdog of the intelligence community is the President’s Foreign Intelligence Advisory Board, created in 1961, with six experienced individuals from outside the government to “conduct a continuing review and assessment” of all intelligence functions. The board, which includes a communications expert among its membership, checks up specifically on N.S.A.

  These multiple points of contact of cryptology with other elements of the American government show how vast and vital and complex an operation cryptology has become. The cost of the operation is enormous. In 1960, the United States was reportedly spending about $380,000,000 a year to maintain the far-flung N.S.A. intercept network and to forward the material to headquarters, and an additional $100,000,000 a year to pay salaries and operating expenses at headquarters. This excludes any additional costs incurred by the separate service agencies. By 1966, the figure had reportedly zoomed to $1 billion a year, probably including the cost of launching satellites to intercept other nations’ messages. This amounted to about two per cent of the 1966 national defense expenditure of $50 billion, and it meant an expenditure of over $15 a year for cryptologic protection for each American family. This is a budget absolutely without precedent in the history of cryptology. It measures the distance cryptology has traveled just since the War and State departments paid out a total of a third of a million dollars over an entire decade for Yardley’s American Black Chamber.

  The figure also indicates the extraordinary value that the government places on cryptologic material—or, more precisely, on the information contained within the armor of cryptography and on the intelligence obtained by cryptanalysis. Great value demands great protection, and the effort to protect has shaped the characteristic external aspect of N.S.A.: imperviousness, blankness, silence, utter security. The efforts to attain this impregnability are not idle, for numerous attempts have been made by both the free and the Communist worlds to penetrate each other’s cryptologic secrets. This cloak-and-dagger work is the “practical cryptanalysis” which the Soviet Union has always engaged in, and which it has pursued with great vigor throughout the Cold War.

  As early as 1946, Soviet agents obtained from cipher clerk Emma Woikin the gists of plaintext telegrams of the Canadian Ministry of External Affairs, and perhaps details of the ministry’s cryptographic systems. Roy A. Rhodes, a married Army sergeant assigned to the motor pool at the American embassy in Moscow, went out drinking with some Russian “mechanics” around Christmas of 1952 and woke up in bed with a girl who later told him she was pregnant; threatened with disclosure of this episode to his wife, Rhodes revealed details of his earlier cryptographic work to the Russians. In 1954, a 27-year-old British ex-soldier named John Clarence was sentenced to five years’ imprisonment for giving the Russians a codeword of “prime importance” for mobilizing Britain’s northeastern air defense.

  The incidents crowd one upon the other. At 7 p.m. on March 5, 1957, Dhanapolo Samarasekara, a Ceylonese, removed what was almost certainly the Ceylonese diplomatic code from the offices of the Ceylonese delegation to the United Nations in New York and delivered it to Vladimir A. Grusha, first secretary of the Soviet U.N. delegation. They met again an hour later and Samarasekara returned the red-bound book to the filing cabinets of the fourth-floor code room. On July 15, Prime Minister S. W. R. D. Bandaranaike told Ceylon’s House of Commons that the code had been replaced “as a precaution.” In 1959, Vadim A. Kirilyuk, a Russian working for the United Nations, urged an American who the Soviets knew from a scholarship application had worked on cryptographic machines to give information on them and to get a job with a vital U.S. agency, presumably N.S.A. The American strung him along for five visits, until, in January, 1960, Kirilyuk was declared persona non grata and sent home.

  The care and thoroughness that Moscow lavishes upon this work, and consequently the importance with which it regards it, is nowhere better shown than in a letter to M.V.D. representatives in Australia. The Russians had been trying for two years to get information about French diplomatic cryptosystems from Mme. Rose-Marie Oilier, a second secretary doing cryptographic work at the French embassy in Canberra, but had made little progress. On January 2, 1952, M.V.D. headquarters wrote:

  In order that we should be able to make a maximum use of Mme. Ollier’s agent capacities, Pakhomov must in the first place ascertain what type of work she carries out at the Embassy, her daily work routine: when she starts work, when is the lunch-hour break, where she lunches, when she finishes work, etc. It is particularly necessary to elucidate all the details connected with the fulfillment of her duties as cipher clerk, namely: in what room is she engaged on cipher work, where the cipher documents are kept, does she have access to the safe, where the ciphers are kept, and does she carry on her person the keys to the safe, etc. It is also absolutely necessary to elucidate, at first orally, the actual technique of the enciphering and deciphering of cables. The elucidation of all these details is necessary to enable us to determine what would be the best way, least liable to exposure, of
effecting the acquisition of deposits of ciphers of her embassy.

  Pakhomov failed to get any information from Mme. Oilier, and was recalled in part because of this. Moscow then transferred the assignment to Vladimir Petrov (adding “ciphers of countries of Anglo-American bloc”), but he did not succeed either. The Australian Commission on Espionage that investigated this and other matters following Petrov’s defection declared that “If they, the M.V.D., could, unknown to the French, get the key to their communications, the security not only of France but of the whole Western world might well be in jeopardy.”

  Strangely, this bitter possibility came partly true in 1954, when the Russians, having failed with France in Australia, succeeded in Paris. Communists in the message center of the French National Defense Committee stole a War Ministry cryptosystem and used it to read orders to the embattled bastion of Dien Bien Phu. This inside information may have contributed to the ultimate capitulation of that fortress, to France’s consequent loss of French Indochina, and so to the miseries and warfare in Laos and Vietnam that, more than a decade later, still plagued the West.

  The effort to get cryptologic material is not limited to the Russians. The Poles photographed Irwin W. (Doc) Scarbeck, 41, second secretary of the American embassy in Warsaw, naked in bed with his 22-year-old Polish mistress, who then urged him “to get for them the cipher.” They also offered him 20,000 zlotys, or $833, for the cryptographic information, but he refused. On Formosa in 1957, organized rioters with axes concentrated on smashing their way into the American embassy code room during anti-American demonstrations. Failing to force the heavy iron door guarding the codes, they hacked through a six-inch concrete wall. The ambassador on the spot said that this would mean “some readjustment” of U.S. codes, but Secretary of State John Foster Dulles later reassured the nation that none had been compromised.