by DAVID KAHN
As a result, Giskes managed to keep NORDPOL going effectively for the unprecedented period of twenty months. No other funkspiel had ever lasted more than three. Agent after agent parachuted into captivity. The beginning of the end came when two agents escaped from the prison at Haaren and made their way to Dutch authorities in Switzerland, where they told the story. Even then the resourceful Giskes staved off defeat by smearing them through a NORDPOL link as having escaped with German aid to infiltrate S.O.E. But with the escape of three more agents on November 23, 1943, the radio game was finally over. Giskes realized this when the messages from England, usually packed with information, came in leached of anything solid.
NORDPOL dragged on for a few more months as both sides fenced for advantage by sending meaningless messages, until the Abwehr decided to end the pointless exercise. Giskes thereupon composed a plaintext message to be sent to the men heading the Dutch section of S.O.E.
To Messrs. Blunt, Bingham & Co., Successors Ltd., London. We understand that you have been endeavoring for some time to do business in Holland without our assistance. We regret this the more since we have acted for so long as your sole representatives in this country, to our mutual satisfaction. Nevertheless we can assure you that, should you be thinking of paying us a visit on the Continent on any extensive scale, we shall give your emissaries the same attention as we have hitherto, and a similarly warm welcome. Hoping to see you.
With malice aforethought, Giskes ordered it passed on April 1, 1944. All ten remaining funkspiel links sent it; four British stations receipted for it as usual; six did not answer the calls. More than two years after it had begun, NORDPOL finally ended.
It had arranged for 190 parachute drops and received 95 of them, containing 30,000 pounds of explosives, 3,000 Sten guns, 5,000 revolvers, 2,000 hand grenades, 75 radio transmitters, 500,000 cartridges, and half a million Dutch guilders in cash—all of which had fallen into German hands. It had captured 54 agents, 47 of whom were shot without trial that autumn at the Mauthausen concentration camp. The great NORDPOL carillon tolled not only their deaths, but those of the hopes of the Allies to establish a viable underground in Holland. It had kept German defenses in the Netherlands intact and untroubled by saboteurs, and had deceived the Allies as to the capabilities of German forces there: The Hague was not liberated until an amazing seven months after D-Day.
It was the worst Allied defeat in the espionage war.
That clandestine struggle bred a surprising variety of systems of clandestine communication. Perhaps the most common one was double transposition. Like the Dutch Resistance groups, the French Maquis often used it. A group led by Gilbert Renault picked a key from 15 to 20 letters long at random from a book held both by the group and its London radio correspondents. The same key served for both transpositions. About half a dozen nulls began and ended each plaintext; the security check often involved inserting a specified letter among these. The encipherer told London what passage had been used as the key by a series of numbers. For example, 05702 01837 would mean page 57, line 2, with 18 letters taken from that line to generate the numerical transposition key, and with the message text 37 groups long. The encipherer then added his personal keynumber to these digits, converted them into two groups of five letters by a homophonic substitution, converted them a second time into two other groups of five letters in case the first pair was garbled in transmission, placed them at prearranged places in the message, and gave it to his radio operator.
Renault’s network used this system from August, 1940, to the summer of 1943, when, while in London, he compiled a code of five-letter groups representing whole phrases in an attempt to reduce his dangerously long radio transmissions. He drew his vocabulary from a study of a commercial code and his network’s messages. He had planned to encipher the codetext by double transposition, but an elderly English cryptologist with bulging eyes, glasses on the end of his nose, and a gold chain across his vest advised him to change his five-letter groups to five-digit ones so that they could be enciphered by one-time pads, which London produced by machine and would give him. The destruction of each sheet after use would prevent the Nazis from reading his back intercepts even if the unused portion of the pad were captured. Renault took the advice. Back in France, he had the satis faction not only of seeing his transmission times drop from half an hour to five minutes but of knowing that his cryptograms were perfectly secure.
Agents of the American Office of Strategic Services also used double transposition. They often keyed their messages with their school songs, using lines 1 and 2 as keys for the first message, lines 1 and 3 for the second, and so on, to prevent the multiple anagramming of two messages with identical keys. One agent, Peter Tompkins, who at 23 was slipped into Rome in January, 1944, to organize a spy ring, employed as keys two passages from Dante that he had memorized but could easily obtain anywhere in Italy if they slipped his mind. Nulls were sprinkled about freely. Interestingly, O.S.S. did not use a security check, but relied instead on “electronic fingerprinting”—taperecording an agent’s fist before he went into the field. Double transposition was simple and quite secure, but it was occasionally solved. A contributory reason may have been such practices as that of the Resistance group codenamed MARCO POLO, which time after time used the word tobacco as a group of nulls in its transposition blocks in (successful) attempts to get the British to supply that necessity in their parachute drops.
The O.S.S. employed a variety of other cryptographic systems for intercommunication among its branches in London, Chungking, Karachi, Burma, North Africa, and other places around the world. It had one-time pads, with random numbers generated by I.B.M. machines, SINGABAS, which it called “Berthas,” M-209S, and strips. Its headquarters message center was in the basement of the O.S.S. administration building at 26th and E streets, North West, in Washington, and was run by New York socialite lawyer John W. Delafield. It included a specialist department of five or six men, headed by bridge expert Alfred Sheinwold, whose task it was to untangle badly garbled messages, to make sure that O.S.S. ciphers were secure and that keys were not being used too long, and to train cryptographers for the branches.
Individual agents sometimes developed systems of their own for special-purpose use within the networks of spies that they recruited. In Rome, for example, Peter Tompkins employed a Vigenère with a tableau based on the 21-letter Italian alphabet (j, k, w, x, y omitted); one keyphrase was AVANTI TORINO. Another was a one-page code of five-letter groups for communication with a group in the mountains near Visso whose radio signal was very weak; the codegroups were so designed that if either the first or the last letter (which were the same) and any of the intervening three letters were received correctly in proper position, the group was known unequivocally. Thus, in the weather section, the first three groups were sereno (“clear”) = TABCT, pioggia (“rain”) = TBCDT, nebbia (“fog”) = TCDET.
The governments-in-exile, who abhorred any semblance of collaboration with the Nazis and who directed their resistance from their headquarters in London, received information from the legations they maintained in neutral and Allied capitals. The Czech counselor in Stockholm, Dr. Vladimir Vanek, sent about 500 messages to his chief between August, 1941, and March, 1942, when the Swedes arrested him on espionage charges, probably under German pressure. Vanek’s messages, sent to the cable address MINIMISE LONDON through the British embassy, were enciphered in a combined transposition-substitution system. It began with a monalphabetic substitution into a numerical alphabet whose first digit always matched the date. Thus, for the 8th of the month, the alphabet was this, with the* used as a word-divider:
The transposition keys were derived from one of the bibles of Czech nationalism, Tomáš Masaryk’s Svetová Revoluce. (His son, incidentally, who was Foreign Minister of the Czech government-in-exile, was the recipient of many messages keyed through his father’s book.) The pages used were selected apparently at random: page 74 for August, page 391 for September, 25 for October, and so on
. For the 8th of September, the eighth line on the page would be taken. The first 18 letters of the line—“pakani profesors tvi … ”—bred a numerical key for a transposition block for the substitution. The result of this transposition was inscribed into a second block whose key was derived from the last 19 letters of the same line (“politicky aneskodilo … ”). The plaintext sent by Vanek on that date was: Sudar.Pachatel*atentatu*na*něm. Munični* vlaky*ve*švedsku*je*německy*konsul*v*malmoe*nolde. Mame*duvěrně*od*noru.Jlnas37, plus 0 as a null. The ciphertext began: 34232 21333 19293 11121 33020 10121….
Sweden’s Arne Beurling solved it when he saw “a shadow of the system” in the cryptograms. He noticed that 0’s, 1’s, 2’s, and 3’s, which stood out because of their much greater frequency, appeared at approximately regular intervals in the cryptograms. He deduced that this resulted from a preliminary encipherment in an alphabet of the type actually used, followed by a first inscription into a transposition block with an even number of columns. This would columnize all the 0’s, 1’s, 2’s, 3’s, and 4’s that formed the first digits of the two-digit substitutes. In the second transposition block, this stack of digits would form a row, and as the approximately equal columns of this block were transcribed, the digits of that row would appear in the cryptogram at approximately equal intervals. Beurling called it one of the best hand systems he ever faced. A linguist disappointed him in not recovering the literal key from the numerical and so culminating the solution by finding the keybook. The solved messages played a role in convicting Vanek as a spy. The Germans also solved the system. By reading Czech exile messages from various capitals, they learned so much about the uprisings in Czechoslovakia in 1944 that they were able to quell them with relative ease.
Most widely used of the spy systems were the jargon codes. A phrase intoned in French, Dutch, Norwegian, or Italian in the imperturbable voice of a British Broadcasting Corporation announcer would detonate explosions at a German radio station, cause machine-gun bullets to rattle in a wild raid, ignite a wooden railway trestle. For these innocent-sounding “personal messages” were, in large measure, signals to Resistance groups for an imminent operation. Peter Tompkins has evoked the atmosphere of a group listening for one of these messages: “Thus, on the night of our first expected drop we sat tensely round the radio listening to the news in Italian, waiting for the announcer slowly to enunciate his special messages: ‘Catherine is waiting by the well,’ ‘The sun will rise at dawn,’ ‘Johnny needs sandals,’ then, like a stab of light, our own special message: ‘William waits for Mary’! The drop was on, due about midnight.”
Other jargon codes acknowledged receipt of messages and information. The MARCO POLO Resistance group, which sent its 20- or 30-page answers to such requests as “What is the diameter of the pebbles on the beaches of France?” and “Provide a calendar of fair days in the towns of France” to England by airplane in unenciphered form because of their length, heard that the information had gotten safely to England when the B.B.C. broadcast the Alexandrine verse ET LE DÉSIR S’ACCROÎT QUAND L’EFFET SE RECULE (“And desire grows when the result recedes”). Still other code messages were in the nature of authenticators. When a Russian engineer came to work with MARCO POLO to help dig out information about Peenemunde, where the Germans had constructed their secret v-weapon launching sites, the B.B.C. confirmed his identity with the phrase LES ÉLÉPHANTS MANGENT LES FRAISES (“The elephants are eating the strawberries”), which is nothing if not unmistakable.
Among the most dramatic of the jargon-code messages were the two that set in motion all across France a massive underground sabotage of German transportation and communications. When the B.B.C. broadcast, in French, “It is hot in Suez,” the Resistance was to put into effect the Green Plan, which called for the sabotaging of railroad tracks and equipment. “The dice are on the table” would institute the Red Plan for the cutting of telephone lines and cables. Throughout France Resistance leaders listened tensely for these code-phrases on their hidden radios. As D-Day approached, the number of messages rose into the scores. Finally, at 6:30 p.m. on June 5, 1944, the two crucial messages were broadcast, followed by hundreds of others, such as “The arrow pierces steel.” And, as the underground leaders mentally interpreted them, they were stunned by a blinding realization that the liberation for which they had worked and waited and hoped throughout four dark years of Nazi oppression was now at hand. For many of them, their hearing of a jargon-code message would remain one of the unforgettable moments of their lives.
The most famous of these jargon-code messages was the one announcing D-Day which the Nazis intercepted, recognized—and ignored.
Abwehr headquarters had discovered that the great Allied invasion of Europe would be signaled to the underground by the first stanza of Paul Verlaine’s melancholy “Chanson d’Automne.” The first half of the stanza, when broadcast on the first or the fifteenth of a month, would warn of the imminence of the Anglo-American invasion. The second half would mean: “The invasion will begin within 48 hours … the count starting at 0000 hours of the day following the transmission.”
In January, 1944, Canaris had passed the details to German intelligence units, with orders for them to listen for the two messages. Among the units that had been straining for months to pick them up was that of Lieutenant Colonel Hellmuth Meyer, intelligence officer for the German 15th Army, whose information came largely from his 30-man interception crew. In their concrete bunker at army headquarters near the Belgian border, filled with sensitive equipment, these experts—each of whom could speak three languages fluently—captured virtually every wisp of radio emanation from Allied sources. They had heard many of the jargon messages, which irritatingly eluded their comprehension. They had even intercepted the calls of military policemen using transmitters in jeeps to direct convoys in England, more than 100 miles away. This information had enabled Meyer to learn the names of many of the outfits preparing for the invasion. Recently, however, these calls had stopped. Radio silence had descended upon England, another bit of evidence that the invasion of the Continent was close at hand. Meyer thereupon redoubled his efforts to pick up the fateful Verlaine message.
On June 1, Sergeant Walter Reichling of Meyer’s team was monitoring the messages in French that followed the 9 p.m. B.B.C. news. “Kindly listen now to a few personal messages,” said the announcer. Reichling switched on a wire-recorder. After a brief pause there followed the first half of the first stanza:
LES SANGLOTS LONGS
DES VIOLONS
DE L’AUTOMNE
(“The long sobs of the violins of autumn”). Reichling rushed to Meyer’s office, and the two listened to the recording. Meyer immediately informed the chief of staff, who alerted his own 15th Army, then teletyped the notification to O.K.W. and telephoned the two German headquarters charged with defending against the invasion. But though the message gave clear warning that the invasion was to be launched within two weeks, at the most, nothing was done about it. O.K.W. thought that one of the invasion headquarters had ordered an alert, and that headquarters thought that the other one—headed by Rommel—had done so. But though Rommel must have known about the message, he apparently discounted it, for on June 4 he left for a much-needed home leave.
Meyer, who knew nothing of this, strained to hear the second half of the stanza. “Its awesome significance overwhelmed Meyer,” wrote Cornelius Ryan. “The defeat of the Allied invasion, the lives of hundreds of thousands of his countrymen, the very existence of his country would depend on the speed with which he and his men monitored the broadcast and alerted the front. Meyer and his men would be ready as never before.”
There were at least 15 other D-Day messages which the Germans intercepted and interpreted. On June 2, for example, O.K.H. teletyped to an invasion defense headquarters detailed information received from the R.S.H.A. about jargon messages. Within three days after hearing MESSIEURS, FAÎTES VOS JEUX (“Gentlemen, place your bets”), L’ÉLECTRICITÉ DATE DU VINGTIÈME SIÈCLE (�
��Electricity dates from the 20th century”), or some other messages, O.K.H. warned, “die Invasion rollen.” But the Germans placed most credence in the Verlaine message.
It came at 9:15 p.m. June 5:
BLESSENT MON COEUR
D’UNE LANGUEUR
MONOTONE
(“Wound my heart with a monotonous languor”). Meyer hurried out of his office with what “was probably the most important message the Germans had intercepted throughout the whole of World War II.” In the dining room of headquarters, he breathlessly told General Hans von Salmuth, the 15th Army commander, who was playing bridge, that the second half of the vital message had arrived. Von Salmuth considered a moment, ordered the 15th Army on full alert, then returned to his cards. “I’m too old a bunny to get excited about this,” he remarked to the other players.
The German 15th Army detects the open-code message to the French Underground that warns that the invasion of Europe will start within 48 hours