THE CODEBREAKERS

by DAVID KAHN

  The Ministry of War, on the other hand, meant well, but outside factors defeated its efforts. In 1910, Major Carder came from France to prepare signal and cryptographic liaison with Russia. The following year he came again with a supersecret codebook, complete with superencipherment, of which only eight copies had been printed by the French Army’s geographic service. Four were retained by the French War and Marine ministries; Cartier smuggled the other four—for the corresponding Russian ministries—across the border in his luggage amid novels and Russian-French dictionaries. Soon thereafter he learned that the shifty-eyed Tartar officer to whom he had delivered the two books for the War Ministry had sold one of them to the Germans.

  This typified the atmosphere of corruption that infected the entire Russian military establishment before World War I. The lover of the young wife of the elderly Minister of War dined with the Kaiser and held five German decorations; the not unnatural suspicions that he was a German spy were later confirmed. Russia had herself scored one of the most spectacular spy coups of modern times when it blackmailed the homosexual Colonel Alfred Redl into betraying the strategic plans that the Austro-Hungarian general staff had drawn up for use in the expected war against the Slavs. Fear of a Russian Redl had kept Colonel Andreiev, in charge of the Army’s cipher bureau, from distributing copies of the new and secret cipher that he had drawn up for war use until the very last minute.

  His caution bred disaster.

  The Russian plan of campaign against Germany in 1914 called for an invasion of East Prussia by two armies. The 1st Army was to drive straight west into that province and grip the German defenders tightly in battle. The 2nd Army, to the south, was to circle around the Masurian Lakes, come up behind the Germans, block their retreat, and destroy them. This strategy naturally required careful timing and close collaboration between the two forces. Unfortunately, Russian communications were woefully inadequate. The 2nd Army had only 350 miles of wire all told to string during its advance across the plains of Poland; this pitiful supply contrasts sharply with the 2,500 miles of wire later used in a single day by an A.E.F. army on the Western Front. At the same time radios were issued only to the headquarters of both armies and the headquarters of their immediate subordinates—their corps. Division and lower headquarters lacked them. The several corps headquarters therefore used their wire to link up with their divisions. Since army headquarters had exhausted their meager wire supplies in stringing lines to the rear commands, this left wireless as the only means of communication among the several corps headquarters and between them and their army headquarters—the two highest echelons of field command.

  Their messages lay naked to the enemy. The general inefficiency that crippled the Russian mobilization had fouled up distribution of the new military cipher and its keys. Within a single army (the 2nd), for instance, the XIII Corps did not have the key needed to read cryptograms from its immediate neighbor, the VI Corps. The war broke out August 4. Before a fortnight had passed Russian signalmen were no longer even trying to encipher messages, but were passing them over the radio in the clear.

  In accordance with the Russian strategy, General Pavel Rennenkampf, commanding the 1st, or northern, Army, began moving into East Prussia on August 17. The German general staff had long foreseen the two-pronged attack—the terrain made it obvious. They had left only one army to defend East Prussia because their strategy called for a quick and decisive victory against France first. This single force was approximately as strong as either Russian army but desperately weaker than both combined, and the general staff had dictated as its strategy to strike with all possible strength at the first Russian force within reach, then to turn and attack the second. East Prussia was the homeland of the Junkers. The Germans preferred not to yield it to the hideous trampling of the Slav.

  They gave battle to Rennenkampf at Gumbinnen. Under a hammering Russian artillery barrage, the German troops broke and fled 15 miles to the rear before they could be halted. The frightened German commander prepared to fall back to the Vistula River and abandon East Prussia. He reported his intentions to the German high command, which promptly began looking for a replacement. But his brilliant First Chief Staff Officer, Colonel Max Hoffmann, pointed out that the southern Russian army had already invaded so far that its left wing was actually closer to the Vistula than the German rear and so was in a position to cut off the German retreat. He convinced his chief that he had to strike against this wing to give the German army freedom to maneuver, if only to reach the safety of the Vistula. The Germans had somewhat mauled the Russian bear before their rout, and Rennenkampf, instead of pursuing to turn victory into triumph, had paused to lick his wounds. Hoffmann was confident that he would rest another day or two. He proposed, and his general agreed, to disengage two German corps from the front against Rennenkampf, switch them southward over the excellent network of German railroads, and fall upon the Russian southern prong with surprise.

  The movement was in its early stages when the new German commander, Paul von Hindenburg, and his chief of staff, Erich Ludendorff, who really ran the show, arrived and confirmed it. The difficult entrainment process began. Ludendorff flung out a screen of cavalry along the northern battle line to conceal the withdrawal of his troops and to keep Rennenkampf under observation. The division of forces violated the German strategic doctrine of concentration, and the question arose as to whether all German forces should be thrown into the battle against the southern force, commanded by General Aleksandr Samsonov. To do so would almost ensure victory, but it would also leave the German rear entirely unprotected from an attack by Rennenkampf. While the German staff was discussing the pros and cons of this move on the evening of August 24, a motorcyclist brought in two Russian intercepts. They had been forwarded on the initiative of the head of the radio station at the German fortress at Königsberg. His operators, who had little traffic of their own to transmit, had begun listening in to the Russian transmissions as a diversion.

  Both messages were from the headquarters of Samsonov’s XIII Corps, which was communicating with army headquarters by radio because that was the only means the corps had. And both were in the clear because XIII Corps had never received the proper cipher key. They specified exactly where the corps was going, when it expected to be there, and what it would do next. Was it a trick? No, because these details were perfectly consistent with an overall Russian directive that had been found in the wallet of a dead Russian officer the day before. The intercepts did not answer the crucial question of Rennenkampf’s intentions. But Ludendorff decided that, with this intelligence, the likelihood of overwhelming victory over Samsonov was worth risking defeat by Rennenkampf. The orders went out to march the remaining troops facing Rennenkampf across the short inner distance between the two pincers.

  The march was getting under way next morning as Ludendorff and Hindenburg appeared at headquarters in Marienburg. But Ludendorff was not entirely free of anxiety about what he had done; second thoughts disturbed him. His thin line of cavalry could have been easily pierced by the Russian 1st Army. “Rennenkampf’s formidable host hung like a threatening thundercloud to the northeast,” he worried. “He need only have closed with us and we should have been beaten.” Their defeat would have meant a tremendous moral blow to the German cause, loss of the country’s richest grain and dairy lands, and possibly the fall of the only barrier between the Russian steamroller and Berlin. Should he perhaps have been a little more cautious? While there was yet time, should he leave some troops to block Rennenkampf? Or should he even call off the whole offensive against Samsonov and turn back against Rennenkampf? So much was at stake, and it rested upon little more than his soldier’s intuition that Rennenkampf would merely crawl forward as he repaired his supply lines and refitted his troops.

  But at headquarters that morning there arrived what at one stroke lifted the burden from the minds of Ludendorff and Hoffmann and permitted them to prepare one of the great military triumphs of the war. It was a Russian intercept. It
, too, was in clear, but this one was from Rennenkampf to his IV Corps, and it read, in part:

  The army will continue its attack. On August 25 it will reach the Wiberln-Saalau-Norkitten-Potauren-Nordenburg line; on August 26 the Damerau-Petersdorf-Wehlau-Allenburg-Gerdauen line.

  Their maps told the Germans that Rennenkampf was still moving at his snail’s pace. The evidence of hasty German departure that the Russian general had seen as he advanced leisurely upon their evacuated positions had confirmed his erroneous opinion that the Germans were in full retreat after Gumbinnen. He did not want to press them too much for fear of forcing them to the Vistula before Samsonov could crush them. The Germans, however, saw at once that he could not reach any position in time to attack the German rear before the expected destruction of Samsonov was complete. Relieved, they concentrated at once on engineering that destruction.

  Later that morning, as the German commanders were returning to headquarters from a conference at a corps headquarters, they stopped at a railway station in Montovo for news. A signalman handed Hoffmann still another Russian intercept—also in clear. Samsonov had sent it to the cipherless XIII Corps at 6 a.m. It was a long dispatch, and Hindenburg and Ludendorff had already driven off when Hoffmann got it all. He sped after them in his own car, overtook them, and, as the two automobiles jounced side by side along the rutted Polish road, handed it over. Hindenburg stopped his car, and the officers studied it:

  … On 25 August the 2nd Army proceeds to the Allenstein-Osterode line; the main strength of the army corps occupies: XIII Corps the Gimmendorf-Kurken line; XV Corps Nadrau-Paulsgut; XXIII Corps Michalken-Gr. Gardienne …. The I Corps to remain in District 5, to protect army’s left flank ….

  It was, in fact, nothing less than a full roundup of the situation as Samsonov saw it, together with the most detailed and explicit moves to be followed by his army. It gave the Germans a knowledge of enemy intentions unprecedented in the whole of military history. It was like reading the mind of a chess opponent, like playing blindman’s buff without the blindfold. It was almost impossible to lose.

  The Germans formulated their plans to take advantage of the weaknesses of the Russian dispositions. They plotted a double envelopment of Samsonov, and it worked to perfection. General combat opened the next day, the 26th. One of the Germans corps marching down from Rennenkampf’s front struck hard at Samsonov’s right; during the night, that wing was turned. Before dawn on the 27th, a hurricane barrage of artillery demoralized the hungry, tired troops of his left flank, and before noon they had fled the field without a single serious German infantry assault. Soon the realization penetrated to Samsonov that instead of the Russians crushing a retreating German Army, that army had in fact almost enveloped him. His XIII and XV Corps, in the center, fought bravely in the confused, surging struggle, but the frantic orders and cries for help that their radios squealed in clear were all heard by the Germans who, fully informed, could exploit a gap here, a movement there. Bit by bit the Germans drove in behind the two corps from both sides; soon the Russians found themselves fighting both front and rear. By the 30th, the Germans had encircled the corps with a ring of steel from which only 2,000 Russians escaped. This ended the battle: there were no Russians left to fight. By then Samsonov was dead. He had shot himself in despair as he and his staff stumbled through the forest in the dark night of defeat.

  Gradually, it became clear to the Germans that they had won, as Hoffmann wrote, “one of the great victories in history.” Almost 100,000 Russians were taken prisoner. An estimated 30,000 were dead or missing. The Russian 2nd Army had ceased to exist. One of the few battles of the entire war that was a decisive victory, Tannenberg—as the Germans named it—demonstrated that the Russian steamroller was not quite the invincible machine that had terrorized central Europe. It catapulted Hindenburg to a popularity that carried him, later in the war, to supreme command, and, in peace, to the presidency of his country. Pro-German groups in Russia began to agitate for a withdrawal from the war. Russian morale sank.

  Hoffmann, the architect of the victory, acknowledged its real cause. “We had an ally that I can only talk about after it is all over—we knew all the enemy’s plans. The Russians sent out their wireless in clear.” The case was clear-cut. Interception of unenciphered communications had awarded the Germans their triumph. Tannenberg, which gave Russia the first push on her long slide into ruin and revolution, was the first battle in the history of the world to be decided by cryptologic failure.

  So inexhaustible were the manpower resources of Russia that not even a debacle like Tannenberg could cripple its war effort. “We are happy to make such sacrifices for our allies,” replied the Grand Duke Nicholas, commander in chief of the Russian armies, when the French ambassador expressed his condolences. And even though the Germans turned on Rennenkampf and drove him out of East Prussia in the Battle of the Masurian Lakes, two Russian armies pounded the Austro-Hungarian forces back through Lvov with such force that they retreated almost to Krakow. Meanwhile, though still plagued with shortages of all kinds, including signal equipment, the Russians finally managed to distribute their cipher system to all commands by the middle of September. On the 14th, the Stavka, the Russian high command, prescribed its use for all military orders.

  The system was a numerical polyalphabetic which negated most of the advantages of polyalphabeticity by enciphering several letters in succession in a single cipher alphabet. It resembled the feeble cipher used by Cornwallis in the American Revolution and solved with ease by James Lovell. Along the top of its tableau were listed 33 letters of the Russian alphabet; the tableau proper consisted of eight lines of two-digit numbers in mixed order. Each line differed from the others, and they were numbered at the left in mixed order. In enciphering, these cipher alphabets were used in rotation, the one numbered 1 first, the one numbered 2 second, and so on. Each alphabet enciphered several letters at a time. The number of letters to be enciphered in a given alphabet before the next came into play lay at the whim of the encipherer, who informed the decipherer of this number by writing it out five times and then placing this group at the head of the cryptogram. If he wished to change this number during a message, he simply repeated the new encipherment group length five times, inserted it into the body of the cryptogram, and used that length from then on.

  Cryptograms in the Russian Army cipher thus consisted of groups of monalphabetically enciphered letters, with the length of the groups clearly indicated by the unmistakable appearance of, say a 99999 (the maximum length) or a 66666. Aside from being vulnerable to the usual techniques of frequency analysis, the cipher would often mirror the telltale repeated-letter pattern of an underlying plaintext word, such as attack or division, that had fallen entirely within a single encipherment group and so had been monalphabetically enciphered. Such a system does not interpose insuperable difficulties to the cryptanalyst, especially when, as with the Russians, it was poorly used, often with intermixture of plaintext. Mixed text was soon prohibited, but by then it was too late.

  For the brilliant young Captain Hermann Pokorny, head of the Russian subsection of the Austro-Hungarian Dechiffrierdienst, had cracked the system and reconstructed all its alphabets by September 19. His first important solution, on September 25, disclosed General Novikov’s lengthy report of his reconnaissance of Central Powers troops, with his additional note: “I took the decision of not crossing the Vistula.” The message was dated 8:40 a.m.; by 4 p.m. the Austrian liaison officer had brought it to the attention of the German headquarters. Knowledge of Novikov’s decision determined the initially successful Austro-German tactics of the battles of the Vistula and San rivers. Other intercepts were valuable in more local situations. A message of Prince Engalitschev, colonel of the 10th Russian Cavalry Division, warned of a strong attack on the fortress of Przemysl; the prepared commander easily warded it off until the Austrian advance forced the Russians to lift the siege in mid-October. During this advance, Pokorny’s group solved as many as 30 cryptograms a day
.

  It was at about that time that the Russians made their first key change. It apparently consisted only of altering the order in which the cipher alphabets were to be used, the alphabets themselves remaining unchanged. Solution of this would have taken Pokorny at most a few minutes. Any difficulty that he might have encountered evaporated when a Russian station repeated in the old key a message already sent in the new.

  Meanwhile, the Germans had, more by fortune than by foresight, developed a cryptanalytic service of their own. Ludwig Deubner, a professor of philology at the University of Königsberg who had enlisted in the Landsturm as an interpreter of Russian and who was stationed at the Konigsberg fortress, began his radio-intelligence work by translating the cleartext intercepts that the fortress radio station picked up. As words in cipher began to appear, he undertook to solve them. Gradually he mastered the Russian system so that he could read messages entirely in cipher. At the end of September, he was called to headquarters and given charge of a group of interpreters who were to learn cryptanalysis. Soon he and an outstanding colleague—Hoffmann called them “quite geniuses in deciphering”—were, with their neophyte code-breakers, sending a stream of solutions to Ludendorff each night about 11. The chief of staff waited for them impatiently, barking, “Any radiograms?” at his subordinates. He based his orders for the next day in large measure on the intelligence the intercepts gave. When they were late, he would stalk into the cryptanalytic section to find out what the delay was. And if for a time nothing of importance appeared in the messages, he would growl that the intercept service had not been paying attention.