by Kahn, David
Bertrand did not reply to Schmidt’s comment, asking him instead, “Are you sure that the documents you have given us are those for the machine currently in service?”
When REX translated the question, Schmidt started. He frowned. He looked at Bertrand, then back at REX.
“I’m not a crook,” he replied. “If you have such doubts, it’s because your cryptographers are incompetent.”
Another French intelligence officer present quickly turned the conversation to the prospect of food, and the matter rested there.
Bertrand asked the question because the French cryptanalysts had not been able to break the Enigma, even with the help of the Schmidt documents. Nor had the Poles reported any successes. Yet they seemed glad to have Schmidt’s keys. Langer and Ciȩżki were friendly to Bertrand. They told him of their intercepts of German messages; they even invited him in 1938 to Pyry.
But they never told him of their reconstruction of the Enigma or showed him a single Enigma solution. Their discretion probably stemmed from a fear that any leak about their reading of German messages would degrade relations with Germany after the signing of the nonaggression declaration. Bertrand perhaps understood this aspect of Poland’s delicate international situation, pressed as she was between two vengeful powers, just as he understood the need for a French ally in the east to threaten Germany from the rear. Apparently, his superiors also understood the situation. France had nothing to lose by supplying keys to the Poles, so Bertrand was allowed to continue doing so. Altogether the French gave the Biuro Szyfrów keys for thirty-eight months out of eighty-one. Surely the French hoped that eventually something would come out of it.
Astoundingly, however, none of these later contributions reached the Polish codebreakers. They were not given a single one of the keys that Bertrand delivered during those five years!
What made the superiors of the codebreakers deliberately withhold documents that would have so lightened and accelerated their travails? The only answer that makes sense is that the chiefs wanted to develop a cryptanalytic capability that would function even if Schmidt quit or was caught, transferred, or turned, or if France, which was irritated at the Polish–German nonaggression pact, ceased supplying his data. Denying the cryptanalysts the keys indeed reduced the volume of Polish intelligence—but increased its independence. And if the cryptanalysts failed to make progress, they could always be given the keys, as was done in 1932, when the bosses finally gave Rejewski the first keys that Schmidt had provided. The withholding succeeded, and even better than the chiefs had hoped. For not only did the codebreakers resolve by pure analysis the fundamental problem of reconstructing the daily keys, not only did they keep up with the successive German security measures, but they continually improved their methods, thereby putting out more solutions faster.
This progress ended abruptly on December 15, 1938. On that day, the Enigma messages became unreadable. The Poles soon learned that the Germans had put into service two additional rotors. Though the Enigma still held only three rotors at a time, these were now chosen from a group of five instead of three, raising the number of rotor choices and orders from six to sixty. And the wiring of the two new rotors was unknown.
Fortunately for the Poles, however, the nets of the Sicherheitsdienst, or SD, the Nazi party intelligence service, had not shifted over to a new indicators method that the army nets had begun to use on September 15. The SD messages enabled the Poles to use their old methods of determining the rotor order and rotor setting as well as the ring positions and the plugboard connections; with this knowledge, Rejewski could reconstruct the wiring of first one, then the other of the two new rotors, as he had done in 1932.
But the new number of rotor orders expanded the task the Poles faced by an order of magnitude. Instead of six bomby, they would need sixty, at a cost that was fifteen times the whole equipment budget of the Biuro Szyfrów for fiscal 1938–39. Instead of 156 Zygalski sheets, BS-4 would need 1,560—and it had by then punched only 52. Nor was there any way out of this, for the Poles could no longer fall back on Schmidt’s supply of Enigma keys. He had transferred on September 28, 1938, to a better job in another communications intelligence agency, Luftwaffe chief Göring’s Forschungsamt. Though he still provided other information for money, he no longer had access to the Enigma keys.
The new situation overwhelmed the heroic capabilities of the Polish cryptanalysts. To make matters worse, on January 1, 1939, the number of plugboard connections rose to ten.
The head of the general staff’s intelligence bureau proposed a meeting with the British and the French in hopes that they would have something to contribute. By chance Bertrand, in France, was then considering the same idea. The meeting took place on Monday and Tuesday, January 9 and 10, 1939, in the French intelligence service headquarters, which was housed in prefabricated structures huddled against the Invalides, the domed tomb of Napoleon; the Service de Renseignements had moved there in 1932. Present were Bertrand; a French cryptanalyst, Captain Henri Braquenié; Langer, the head of the Biuro Szyfrów; Ciȩżki, the head of BS-4; Alastair Denniston, the head of Britain’s codebreakers, who had been one of the original members of Room 40; and two other Britons. The atmosphere was cordial, but no one revealed any solutions of the machine: the British and French because they had none, the Poles because they had been instructed to say nothing unless they got something in return. The conclusion of the conference was pessimistic:
Reconstruction of the machine solely by the study of enciphered texts is proving itself practically impossible; this was, furthermore, the view of the German specialists who devoted themselves to the same task before putting their machine into service…. The labors undertaken [by the three nations] seem to have ended in an impasse from which only information from an agent will enable them to escape, so a technical questionnaire, as simple as possible, has been drawn up, to be given to an agent judged able to carry out such an assignment.
Two months later, Hitler, who had said at Munich that he wanted only the German-speaking parts of Czechoslovakia, occupied the rest of that country. The scales fell from the eyes of the British and the French. On March 30, Britain, seconded by France, guaranteed Poland “all support” in the event of an unprovoked attack by Germany and six days later signed a provisional treaty of assistance with Poland. On April 27, Germany, blaming that agreement, declared that its 1934 nonaggression agreement with Poland was “null and void.” This declaration eliminated one of Poland’s chief reasons for not revealing her cryptanalytic successes to France and Britain: fear that a leak would provoke Germany into denouncing the agreement, with all the bullying that might follow.
Then, in May, two developments impelled Poland to share the results of her codebreaking. First, in a secret agreement, France promised to advance with the bulk of her forces against Germany fifteen days after any German invasion of Poland, and Britain began military talks in Warsaw. Second, tension between Poland and Germany approached the breaking point. Hitler’s anti-Polish speeches incited Germans in both countries; the fatal shooting of a German by someone in an official Polish car became a cause célèbre; Nazis marched in the Free City of Danzig, and Nazi propaganda minister Joseph Goebbels warned that no power on earth could prevent the return of that city to the Reich; Poles and Germans fought in Polish Silesia; bombs exploded in Polish homes; a German plane was downed over a Polish naval base; the Reich press reported attacks on Germans in Poland.
It was in this atmosphere that Langer, on June 30, telegraphed London and Paris. He said that something new had come up since the January meeting and invited the French and the British to another conference, this one in Warsaw.
On the morning of Monday, July 24, Bertrand and the cryptologist Braquenié arrived in Warsaw by the overnight Nordexpress, which linked Paris and Moscow. Three Britons flew in: Denniston, Dillwyn Knox, and Commander Humphrey Sandwith, the head of the section that had developed and controlled the Royal Navy’s intercept and direction-finding stations. The French wer
e put up at the Hotel Polonia, and the British at the Bristol, where Mayer, the head of intelligence, Langer, Ciȩżki, and the three young cryptologists entertained the visitors at lunch. The conversation, which dealt only with banalities, was, ironically, in German, the only common language of the Poles, the Englishmen, and Bertrand.
Later all the cryptologists traveled out to Pyry. Langer gave them a short tour of the cryptologic center and then took them into the cryptologists’ office. On tables stood several objects under covers. When all had gathered around, Langer, without a word, removed the covers, revealing the Polish replicas of the Enigmas, which the French and British cryptologists recognized at once. They were utterly astonished.
“Where did you get these?” Bertrand asked.
“We made them ourselves,” Langer replied.
The Britons examined a machine closely. Sandwith threw Langer an incredulous glance, and Langer repeated that the Polish cryptologists had built it. Knox, who was intimately familiar with the Enigma from his own work on it, asked the most questions.
One question was, what was the wiring from the keyboard to the first rotor? This problem, which Rejewski had solved in a flash of intuition, had frustrated Knox. He felt cheated when he was told that it was Q to Q, W to W, E to E, and so on. Denniston wanted to ring up the British embassy to have London send draftsmen and electricians to draw up the plans of the Polish machines so that they could be reproduced. Langer restrained him: more was to come.
In the next room, the foreign cryptologists saw six cupboards about four feet high: the bomby. Langer demonstrated how they worked. Rejewski answered questions, since the machines had been built to his design. Zygalski explained how his perforated sheets worked. The Poles described the various methods for recovering keys; they explained that for a long time, using the bomby, they had been able to find an Enigma key, under the right conditions, in two hours. Then the Germans had placed two additional rotors in service. Cryptanalysis now required ten times as much equipment as before if solutions were to be anywhere near current, and the Poles’ capabilities had been outstripped. France and Britain had greater resources. Poland’s sharing of her cryptanalytic knowledge with her two allies would enable them to continue solving the German cryptograms. This would be Poland’s contribution to combating the common German menace.
The Britons and the French expressed their delighted thanks. Denniston again sought to telephone for the technicians. He could hardly believe his ears when Langer told him that the Poles had prepared two Enigma replicas for their visitors: one for the French, one for the British. These would be given to Bertrand for shipment under diplomatic seal to Paris and then for forwarding to Britain. The conference ended on Tuesday in an atmosphere of warmth, astonishment, gratitude, and anticipation.
A few days later, Langer, on his way to Britain, passed through Paris. Bertrand and one of his superiors took the Polish cryptologic chief to lunch at Drouant, a classic Paris restaurant, where they feted him with champagne. “We owed him at least that!” said Bertrand. Soon afterward, the machines arrived. The equipment destined for Britain was packed into the largest diplomatic bags available at the embassy and placed under British diplomatic seal. The British intelligence liaison officer to the French, Dunderdale, and three assistants brought it to Paris’ St. Lazare railroad station. Bertrand joined them on the boat train, the Golden Arrow. It was August 16. In Germany, troops were readying an attack on Poland. Captain Karl Dönitz, the submarine commander, had ordered his U-boats to their war stations at sea.
At Dover customs in England, Dunderdale ran into the French playwright Sacha Guitry and his wife, the actress Yvonne Printemps, on their way to the opening of a play of Guitry’s in London. Dunderdale knew them by sight since they lived opposite him in Paris. Their luggage was voluminous, and Dunderdale made a deal with them: he would get it through customs duty-free if they would pretend that the bulky diplomatic baggage was theirs, to throw off the suspicions of any spies. The couple agreed and were waved through; the Enigma bags drew the attention of no German agents. That evening, the Golden Arrow pulled into London’s Victoria Station near the end of the rush hour. The deputy chief of the British foreign intelligence service, Colonel Stewart Menzies, on his way to a reception, awaited them in black tie. “Accueil triomphal!” (Triumphal welcome), thought Bertrand. Menzies’s men carted away the precious mechanism.
Once again, as in World War I, Britain had obtained the means for solving her adversary’s messages through a gift from a loyal ally.
6
FAILURE AT BROADWAY BUILDINGS
EARLY IN 1919, PROFITING FROM THE LESSONS OF THE GREAT WAR, the British cabinet had decided to establish a permanent codebreaking agency. The first lord of the Admiralty, who was both interested in intelligence and politically powerful, captured the agency for the navy, perhaps in part because it ran most of the intercept stations that provided the cryptanalysts with raw material. The director of naval intelligence, Captain Hugh (Quex) Sinclair, began by recruiting veterans of the navy’s Room 40 and the army’s M.I.1b. In particular, he brought in the two brightest lights of those bodies, Room 40’s Dillwyn Knox, and M.I.1b’s Oliver Strachey (the older brother of the Lytton Strachey who, years before, had tried to seduce Knox). Sinclair put Alastair Denniston in charge. Denniston, known because of his stature as “the little man” to his subordinates, was not the best administrator: one person spoke of him viciously “as possibly fit to manage a small sweet shop in the East End.” But Denniston’s dislike of hierarchy protected his individualistic cryptanalysts from the rigidities of bureaucracy. And his knowledge of the subject matter, his informality, and his willingness to delegate responsibility helped him build and sustain a secret government department of great value.
The new organization, which had the public function of securing the government’s communications and the secret one of reading foreign governments’ messages, was given the deliberately misleading name, proposed by a Foreign Office staffer, of Government Code & Cypher School. It officially came into being November 1, 1919, with 66 staffers—29 professional and 37 clerical, 12 of these for constructing British codes. A year and a half later the new foreign secretary, Earl Curzon of Kedleston, who had long wanted control of codebreaking, took advantage of a political leak of some Soviet solutions and of the fact that the first lord of the Admiralty had been replaced by a man who had neither his political clout nor his understanding of intelligence and who was glad to reduce costs. By April 1921, G.C.&C.S. was under the practical control of the Foreign Office, and by April 1, 1922, under its formal control as well.
The next year the universally admired Quex Sinclair rose from director of naval intelligence to chief of the Secret Intelligence Service (S.I.S.)—“C,” in the official hieroglyph. The Foreign Office agreed to let him assign G.C.&C.S. its work while it administered the agency. Soon after he took over, G.C.&C.S., which had been housed first behind the Charing Cross railroad station and next in distant quarters in Kensington, moved in 1925 into the third and fourth floors of the ten-story gray concrete Broadway Buildings. The office building, at 54 Broadway, a couple of blocks from Westminster Abbey, and housing as well the S.I.S., was a few hundred yards from the Foreign Office, which codebreakers visited an average of eight times a day, and only a little farther from the Admiralty. G.C.&C.S. expanded by 1935 to about 90 employees, 30 of them cryptologists, and by mid-1939 to 200, with 33 cryptologists. In addition, between 140 and 240 servicemen were intercepting foreign transmissions. Thus, by 1939, almost 500 persons worked in British cryptology at an annual salary cost of £100,000 (or about $3 million in 1991 dollars).
The armed services provided not only the intercept personnel but the intercept posts themselves. Economics compelled cooperation. The War Office handled the Middle East; the Admiralty, the Far East; the Air Ministry picked up what it could hear within the United Kingdom. The need to centralize interception, traffic analysis, and cryptanalysis under one roof largely withstood the centrifug
al pulls of the different services to withdraw elements from G.C.&C.S. or establish their own duplicative bodies.
Like all agencies, G.C.&C.S. continually reorganized itself. Early in 1924 one of the more important of the Room 40 recruits, the prickly lawyer William F. (Nobby) Clarke, whose principal interest was naval cryptanalysis, was on vacation near Nice when a suicide in G.C.&C.S. opened up a senior assistant’s slot. Denniston promoted one of his prewar friends who had been a subordinate of Clarke’s in Room 40 and whose work, Clarke thought, although good, had not been outstanding. Clarke was furious. Denniston explained upon Clarke’s return that Sinclair had not chosen Clarke because Clarke was too busy on Admiralty assignments, among them lecturing to naval officers on cryptology to impress upon them the lessons of the war. The codebreaker protested violently that if his naval work was prejudicing any promotion, he wanted to be relieved of it. This put Denniston in a difficult position, as there was no one else capable of doing it. To Clarke’s surprise, Sinclair invited him to a private lunch at his home.
No junior [assistant] had ever had that pleasure, pleasure it was, for he was an excellent host [Clarke wrote]. I was the only guest, another change, and it was soon clear that I had been asked because he was anxious to hear my views. These I gave quite frankly and emphasized the importance of the naval side of our work and the vital necessity, in my opinion, of having a proper naval section; he listened, as he always did, most attentively, asking his usual searching questions and said he would think the matter over. Soon afterwards Denniston saw me, said the formation of a naval section had been decided on, that I was to be its head and that I would be promoted at once. This duly took place about the middle of 1924.