THE CODEBREAKERS

by DAVID KAHN

  Meanwhile, Meyer telephoned his headquarters, and later dispatched “Teletype No. 2117⁄26 urgent to 67th, 81st, 82nd, 89th Corps; Military Governor Belgium and Northern France; Army Group B; 16th Flak Division; Admiral Channel Coast; Luftwaffe Belgium and Northern France. Message of B.B.C., 2115, June 5, has been processed. According to our available records it means ‘Expect invasion within 48 hours, starting 0000, June 6.’” It was then only about three hours until 18,000 paratroopers would drop into the hedgerows of Normandy to begin the invasion, only about eight hours to the H-Hour landing on Omaha, Juno, Sword, and the other beaches. All the German commands were notified—except one. For reasons that have never been fully explained, the German 7th Army, the one on which would fall the brunt of the cross-channel attack, was never alerted to D-Day.

  In the United States, a great battle to wrest almost infinite power from the infinitesimal atom was plunged in secrecy as deep as nature’s own. Congress did not know about it: the two-billion-dollar cost of creating the atomic bomb was appropriated from a special presidential emergency fund. References to uranium were kept out of the newspapers. The huge plants and laboratories necessary were built in sections of the country as remote as possible—Oak Ridge, Tennessee; Hanford, Washington; atop a mesa at Los Alamos, New Mexico. And everybody and everything had a codename, beginning with the very name of the project: the MANHATTAN ENGINEERING DISTRICT. Even before it came into being, bomb was being called THE GADGET, THE DEVICE, S-1, THE THING, THE BEAST, or simply IT. Later, after its probable dimensions had evolved, the scientists called the uranium bomb, which would bring the two masses of uranium into one of the critical size by shooting one into another down a gun barrel, the THIN MAN after President Roosevelt. The plutonium bomb, which would implode a sphere of plutonium and thus would require a bulkier casing, was called, in contradistinction, the FAT MAN after Churchill. When the THIN MAN’s gun barrel was shortened, it became known as the LITTLE BOY.

  The man in charge of the project, Brigadier General Leslie R. Groves, was sometimes called RELIEF, sometimes 99—from the way his secretary wrote “G.G.” for “General Groves.” In fact, some of the scientists who were less bemused than most by the secret-club nature of the project simply referred to him as “G.G.” A special operation to ascertain German atomic capabilities was called ALSOS, a Greek word meaning “groves.” Dr. Arthur Holly Compton became A. H. COMAS or A. HOLLY. Niels Bohr, one of the top atomic scientists, was rechristened NICHOLAS BAKER, and Enrico Fermi became HENRY FARMER. The Los Alamos laboratory was SITE Y; K-25 was the gaseous diffusion plant at Oak Ridge and Y-12 the electromagnetic plant there. For Manhattan District purposes, the University of Chicago was known as the CHICAGO METALLURGICAL LABORATORY, and the first man to take charge of the atom-splitting work there, Gregory Breit, was rather uninhibitedly referred to as the COORDINATOR OF RAPID RUPTURE.

  Though telegraphic messages were handled by the Signal Corps with its own cryptographic equipment, a need arose for secrecy in telephone conversations in the nationwide project, which frequently sent key officials into the field far from any cipher machines. The first telephonic cipher was devised on an impromptu basis by Groves’s secretary, Mrs. Jean O’Leary, when she had to telephone some secret information to him. “Go get what you always see me using,” she said. Groves bought a pack of her favorite cigarettes and spelled out the message as she gave each letter by its position in the words on the pack. A few days later, Groves himself made up what he called a “quadratic code” for use with a “number of people with whom I might have had to talk over the phone on matters of high secrecy. Each one was different and as far as I can recall I was the only one who had all the codes. Possibly Mrs. O’Leary kept them all in one of the top-secret safes. People were instructed to carry them in their billfolds and report instantly if they were lost.” The quadratic code was actually a 10 × 10 checkerboard, like this one, a typewritten square about 3½ × 4 inches that Groves used with Lieutenant Colonel Peer da Silva, chief of security at Los Alamos:

  Teller might be enciphered 93, 31, 64, 28, 07, 70, and U-235 as 23, 80, 43, 84, 77. No key pattern is apparent, and the square provides a satisfyingly sufficient complement of plaintext letters to suppress frequency: nine e’s, seven t’s, six i’s, six o’s, and so on. All letters are represented. “The codes were not used to spell out an entire message,” Groves has written, “but rather one or, at most, a few key words…. Even in spelling out the key words, it was not usual to spell them out in their entirety. Between [J. Robert] Oppenheimer and myself, only the first two or three letters would normally be needed.” A checkerboard is hardly a high-security cipher, but in view of this brevity, of Groves’s assurance that “all codes were constantly changed, as we recognized how easily any code could be broken if enough messages were available,” and of a spy’s difficulty in tapping just the right telephone wire to intercept one of these messages, it seems adequate to its purpose. In any event, there is no record of anything having been compromised through its interpretation.

  “Most of our telephone calls included a great deal of double talk and references to things and individuals which no one else would easily spot,” Groves has noted. One of these spur-of-the-moment jargon codes was employed by Arthur Compton when, on December 2, 1942, he telephoned James B. Conant, president of Harvard, to report on Enrico Fermi’s unexpectedly early success in producing man’s first controlled chain reaction in a squash court at Stagg Field, Chicago.

  “Jim, the Italian navigator has just landed in the new world,” said Compton, who composed one of the more felicitous jargon expressions in history when he equated atomic fission with a new world and made Fermi the Christopher Columbus of nuclear physics. “The earth was not as large as he had estimated,” Compton continued, indicating to Conant that the size of the atomic pile was not as large as originally thought necessary, “and he arrived sooner than expected.”

  “Were the natives friendly?” asked Conant in oblique reference to possible problems.

  “Yes. Everyone landed safe and happy.”

  This quick resort to indirection sometimes partook, in the heavily intellectual Manhattan District atmosphere, of a somewhat exotic nature. Compton once had to discuss the problem that the by-products of fissionable material might pollute the water supply. He succeeded in conveying this message about this most modern of difficulties through fairly obscure references to ancient Greek mythology. At the other end of the long-distance wire, Crawford H. Greenewalt of du Pont not only understood but replied in kind.

  When the first atomic bomb was exploded near Alamogordo, New Mexico, the test was given perhaps the least felicitous codename in history: TRINITY. Details of its success were reported to Secretary of War Stimson, then at the Potsdam Conference, in still another informal jargon code. To give Stimson some idea of the immensity of the blast, the originator of the message, special consultant George L. Harrison, related its visibility to the 250-mile distance between Washington and Stimson’s Highhold estate on Long Island and its thunderousness to the 60 miles between Washington and Harrison’s farm at Upperville, Virginia. The LITTLE BOY in his message referred, not to the uranium bomb, but to the plutonium bomb just detonated; his BIG BROTHER referred to the uranium bomb that all Manhattan District people now felt confident would explode. DOCTOR harkened back to a previous telegram, in which Groves was referred to as a physician and the TRINITY test as an operation. Harrison’s message read: DOCTOR HAS JUST RETURNED MOST ENTHUSIASTIC AND CONFIDENT THAT THE LITTLE BOY IS AS HUSKY AS HIS BIG BROTHER. THE LIGHT IN HIS EYES DISCERNIBLE FROM HERE TO HIGHHOLD AND I COULD HAVE HEARD HIS SCREAMS FROM HERE TO MY FARM. Stimson understood it all without trouble and interpreted it to President Harry S Truman. But the young signal officers who deciphered it at the Potsdam message center had no inkling of its real meaning, and they speculated irreverently on whether the 77-year-old Stimson had become a father and whether the Big Three would adjourn for a day in respectful tribute.

  The mission of dropping
the atomic bomb on Japan was given the code-name CENTERBOARD, but so secret was the project that the officer in charge of assigning codenames was not told what it was for. The transportation of uranium to Tinian, the Pacific Island where the bomb would be finally assembled and from which the plane would take off for the strike, was referred to as BOWERY shipments. Those who had codenamed the atomic bomb LITTLE BOY had severely underrated it: LITTLE BOY stood 14 feet tall, measured five feet in diameter, and weighed just under five tons.

  When the Enola Gay took off toward Hiroshima on that meteorologically beautiful morning of August 6, 1945, with LITTLE BOY in her bomb bay, she carried also a special code for reporting the bomb’s effects. It was in the hands of Captain William S. Parsons, the atomic expert who flew the mission. He had made it up two days earlier with Brigadier General Thomas F. Farrell, Groves’ personal representative on Tinian. It consisted of 28 items covering every eventuality that they could think of for the drop. Each item had been placed on a separate line, and to transmit the strike report all that Parsons had to do was to read off the number of the appropriate line. In addition, the code was divided into three sections, ABLE, BAKER, and CEDAR, for good, medium, and bad results. One of these words was to be transmitted first as a general indication of what was to follow. Farrell had practically memorized it:

  ABLE Line 1. Clearcut, successful in all respects.

  2. Visible effects greater than TRINITY.

  3. Visible effects about equal to TRINITY.

  … …

  6. HO [Hiroshima, the primary target].

  7. KQ [Kokura, a secondary target].

  8. NA [Nagasaki, a secondary target].

  9. Conditions normal in airplane following delivery, proceeding to regular base.

  … …

  BAKER Line 11. Technically successful but other factors involved make conference necessary before taking further steps.

  12. Doubt as to whether delivery was made to planned target.

  13. Apparent functioning in an unprofitable portion of target.

  … …

  CEDAR Line 21. Apparent technical failure.

  22. Returning with unit to indicated place due to weather.

  … …

  27. Iwo Jima.

  28. Destination.

  At 8:16 a.m., the incongruously named LITTLE BOY, the most awesome weapon that mankind had ever built, obliterated Hiroshima in a blast of fire and destruction that killed 60,000 Japanese, made the very name of the city a synonym for horror, and changed the face of war. A few minutes later, as the Enola Gay headed back to Tinian, Parsons reported the bare facts of the holocaust in a brief code message: ABLE, LINE 1, LINE 2, LINE 6, LINE 9. Farrell translated it aloud to a small group of observers without looking at his code-sheet, and immediately passed the word to Washington. Sixteen hours later, a stunned world learned that it had entered the atomic age.

  Telephoning is an exceedingly convenient way to communicate. How delightfully simple to pick up a phone, talk with the other party, and get everything settled in one conversation! Much easier than sending written messages back and forth. But the telephone is notoriously insecure—and its offspring, the radiotelephone, even more so. A single wiretap grants access to a telephone conversation, and only a radio set is needed to overhear radiotelephone talk. And the Axis did not hesitate to grasp these opportunities at the highest diplomatic levels.

  The most obvious protective measure against eavesdropping is to make up codes for conversation, and this has of course been done at one time or another by almost anyone who has spoken over the telephone. The codes range from mere oblique references and the most impromptu cant to elaborately prepared lists of jargon. Less frequently, a message might be enciphered in a prearranged system and the ciphertext read off letter by letter, as with the Manhattan District checkerboard. Or the speakers may resort to a foreign language.

  The United States raised the latter device to the level of a full-scale system in both World Wars by making use of a resource that virtually no other combatant had: pools of tongues so recondite that almost no one else in the world understood them. These were the American Indian languages, which are isolated both geographically and linguistically. In 1918, eight Choctaws of Company D, 141st Infantry, transmitted orders by field telephone; this was the idea of Captain E. W. Horner, who named Solomon Lewis as the chief of the detail. Other Indian tongues were also used. During preparations for World War II, the Signal Corps tested Comanches and Indians from Michigan and Wisconsin in war games, but most of the codetalkers in the combat itself were Navaho. One reason probably was that the tribe was large enough (more than 50,000 persons) to furnish a goodly number of speakers; another, that reportedly only 28 non-Navahos—mainly anthropologists and missionaries—could speak the language, and none of these were German or Japanese; a third reason was the extreme difficulty of the tongue and the near impossibility—even if someone did learn it—of counterfeiting its sounds.

  “Sounds [in Navaho] must be reproduced with pedantic neatness … almost as if a robot were talking,” wrote anthropologist Clyde Kluckhohn. “The talk of those who have learned Navaho as adults always has a flabby quality to the Navaho ear. They neglect a slight hesitation a fraction of a second before uttering the stem of the word.” A hint of its complexity may be seen in some of its verb forms, on which it insists. The stems of many Navaho verbs differ with the object acted upon. Thus one stem must be used with long objects (pencils, sticks), another with slender flexible objects (snakes, thongs), and still others with granular masses (sugar, salt), things bundled up (hay, bundles of clothing), fabrics (paper, blankets), viscous objects (mud, feces), bulky round objects, container-and-contents, animate objects, and so forth. An entirely different verb form concerns itself with the manner of knowing an event. For example, a Navaho must use one form if he himself is aware of the actual start of rain, another if he believes that rain was falling for some time in his locality before he noticed it, and so on. “Because so much is expressed and implied by the few syllables that make up a single verb form, the Navaho verb is like a tiny imagist poem.” Thus “ná’íldil” means “You are accustomed to eat plural separable objects one at a time.”

  A cryptosystem like that boasts considerable security, and it is not surprising that the dark-skinned, black-haired Navaho became a familiar sight in Marine regimental, divisional, or corps command posts, translating a message into a conglomeration of Navaho, American slang, and military terminology as he huddled over a radio set in the Pacific combat zone. Close friends usually worked together. The number of Navaho codetalkers in the Marines rose from 30 at the start of the war to 420 at the end. They relayed operational orders with a secrecy that helped the United States advance from the Solomons to Okinawa.

  Linguistic codetalking, jargon codes, or double meanings all use the human speaker as the coding machine. But this job may be delegated to a real machine—the scrambler. These two modes of oral secrecy, the human and the mechanical, correspond to the two basic forms of cryptosystems. Human coding transmutes words, syllables, and sounds (as in Pig Latin)—the linguistic elements of speech—into secret forms and so parallels code. Both ciphers and scramblers, on the other hand, work upon particles of a text cut up without regard to linguistic functions. From this analogy, scrambler methods of modifying speech are called “ciphony” (from “cipher” plus “telephony”). The field of secret voice communication as a whole may be termed “cryptophony.”

  Though it was only in World War II that scramblers came into widespread use, and only in that war that serious attempts began to be made to solve scrambled speech, devices to assure telephonic secrecy had been in existence almost as long as the telephone itself. The granddaddy of these was patented on December 20, 1881, only five years after Bell obtained his patent on the telephone. Its inventor, 25-year-old James Harris Rogers, an American electrical pioneer who was then chief electrician for the Capitol, wrote: “My invention consists in throwing a message sent from any tran
smitting instrument through two or more circuits alternately in rapid succession … in such a manner that anyone tapping but one of the circuits is unable to obtain anything but a confused and unintelligible series of signals…. The two or more lines on which a signal is transmitted according to my plan may be carried to a common terminus by widely different routes, and thus it will be impossible for any person wishing to do so to … or tap both lines at the same time.”

  Later methods operate more directly on the speech itself, often in ways that resemble transposition, substitution, and null ciphers. In most of the substitution systems, ciphony selects one component out of the many that make up the complex phenomenon of speech and alters it. It usually chooses frequency, though some scramblers distort volume. Frequency here refers to the number of times the vocal chords vibrate; it is usually stated in terms of cycles per second, or c.p.s., so that a frequency of 500 c.p.s. means that the vocal chords are vibrating 500 times a second. Because of the resonance of the vocal organs, most sounds in speech combine several frequencies, and each sound has its distinctive combination of frequencies. The main frequency of the /ē/ sound in “feel,” for example, is much higher than that of the /ü/ sound in “fool.” Naturally, the absolute frequencies will differ somewhat from person to person, but it is the relative variations within an individual’s speech that carry much of its information content.

  Ciphony seeks to conceal this content by shifting the frequencies of the sounds of speech. It can do this because the telephone first converts these sounds into a fluctuating electrical current, which the tubes, switches, filters, and circuits that comprise a scrambler then modify according to well-known principles of electricity.* Though this current may be transformed in a great variety of ways, many affect the voice essentially alike, so that there are relatively few basic scrambles.

  The simplest is inversion. This turns the voice upside down. Though normal speech ranges from about 70 to about 7,000 c.p.s., the telephone, for engineering reasons, responds only to sounds from about 300 to 3,300 c.p.s. It is this frequency band that is inverted. A voice tone of 300 c.p.s. will emerge from the inverter at 3,300 c.p.s., and vice versa. A tone of 750 c.p.s. will become 2,250 c.p.s., and again vice versa. It is the equivalent of a = Z, b = Y, …, z = A, a phonetic atbash. Inverted speech sounds like a thin high-pitched squawking, ringing with bell-like chimes. The word company resembles CRINKANOPE, Chicago, SIKAYBEE. The inversion pivots in the middle of the frequency band, which means that tones in this area somersault through a narrow range. A frequency of 1,625 will become 1,675. This relative lack of change results in the phenomenon that the word inverter itself, which is composed largely of such tones, emerges from the enciphering process that it describes almost unchanged!