Over the next week and a half three more U-boats were sunk by escorts in depth charge and ramming attacks. Otto Kretschmer of U-99, Dönitz’s top ace, was taken prisoner; Joachim Schepke of U-100, the next highest in the tonnage score, was crushed to death when the destroyer Vanoc sliced through her conning tower, ramming the surfaced boat as she tried to flee in the darkness; Günther Prien and U-47 were lost either in a depth charge attack or from some other cause. The string of British successes had to be partly coincidence, but they were a harbinger that the tide of the battle was shifting at last.
Just after noon on May 7, 1941, a Royal Navy task force of three cruisers and four destroyers reached a bleak stretch of ocean northeast of Iceland, changed course to the east, and formed into seven parallel lines spaced at ten-mile intervals. For the rest of the day they steamed steadily eastward, combing a square of sea sixty-nine miles on a side, which was one degree of latitude. Harry Hinsley, the young German history student who had become Bletchley Park’s authority on German naval communications, had determined that somewhere within that patch of ocean, designated on German charts as grid square AE 39, was the German weather trawler München. Aboard her, Hinsley fervently hoped, were the maddeningly elusive pieces of the puzzle that would allow the Bletchley code breakers to crack the naval Enigma once and for all.
In the fading afternoon light a lookout aboard the cruiser Edinburgh spotted smoke on the starboard bow. After a short chase two of the destroyers overhauled the German trawler and sent a few shells hurtling into the water near her; the crew promptly abandoned ship and a boarding party from the destroyer Somali, which had meanwhile raced alongside at top speed, leapt aboard. They were soon joined by an officer from the Edinburgh: Jasper Haines, a captain in the Naval Intelligence Division. He seemed to know exactly what he was looking for, the other members of the boarding party later recalled, making his way directly belowships and returning a few minutes later with a bundle of papers in his hand. Within a few hours Haines was heading for Scapa aboard the destroyer Nestor, with orders to get to Bletchley as soon as possible with his finds.33
Two days later an unplanned bonus was added to Hinsley’s haul. The destroyer Bulldog was escorting a westbound convoy of forty ships when two enormous torpedo blasts suddenly erupted in the bright noon. The Bulldog’s captain, Commander A. J. Baker-Cresswell, sheered off from the head of the column with two corvettes to chase down the attacker. Almost at once the corvette Aubretia made asdic contact and began dropping depth charges. Only moments after having turned and dropped a second pattern of depth charges, the escorts were rewarded with “the dream of all escort vessels,” as Sublieutenant David Balme of the Bulldog put it, the sight of a U-boat blown to the surface. Thinking quickly, Baker-Cresswell saw the chance to capture the boat before her crew could scuttle her, which was the standard German practice. He ordered his crew to open fire with every available weapon. A hail of shells and bullets from the ship’s two 4.7-inch guns, machine guns, even an antiaircraft gun clattered in deafening syncopation against the metal hull of the U-boat. The German crew began spilling out onto the deck and into the water.
Baker-Cresswell turned to Balme. “Right, we will board her. Sub, you take this sea boat.” Balme was twenty years old, the sea boat was a rowboat propelled by five oarsmen, and the sea was running with six-foot rollers. Reaching the U-boat, Balme climbed into the conning tower, waved his revolver around a bit wondering apprehensively if any Germans were still down there, and plunged down the ladder. The boat was deserted. Balme formed the rest of the boarding party into a human chain and passed charts, books, papers, and gear up the ladder and into the sea boat. The captured boat, U-110, was taken under tow by Bulldog but sank the next morning (“one of the greatest blessings in disguise,” Balme later understood, as it kept the Germans from learning that the British had been able to retrieve the critical documents from the boat).34
On May 13 a Royal Navy intelligence officer met the returning ship at Scapa Flow, and that same day he was on a plane for London, and then on to Bletchley Park, where he arrived that afternoon. That night at 9:37 the teleprinters at the Admiralty’s bombproof concrete “Citadel” in Whitehall came rattling to life with the start of an uninterrupted stream of a hundred deciphered naval Enigma messages from the Bletchley code breakers.35
THE BREAKTHROUGH HAD COME from a list of daily settings for the naval Enigma network taken from the München and, even more priceless, the external code tables used to encrypt the eight-letter indicators which specified at the start of each message the precise starting position of the rotors used for the machine for that particular message. Those tables were among the haul of papers from U-110. (Balme had also retrieved an intact Enigma machine, but that added nothing the code breakers had not already long had in hand.)
Since the previous summer the Bletchley cryptanalysts had been breaking army and air force Enigma traffic using the first of the behemothlike mechanical calculators they called the “bombes.” Built by the British Tabulating Machine Company, they cost £5,000 apiece, weighed one ton, emitted a perpetual stream of leaking machine oil, and were plagued with temperamental wire contacts on the dozens of spinning rotors which reproduced the Enigma machine’s wiring. Each little wire brush contact had to be preened with tweezers before each run of the machine to keep them from shorting.36
They were a stroke of mathematical if not exactly mechanical genius. Alan Turing had worked it out in a flash of jaw-dropping insight. The bombes operated on the principle he had discovered, that if one knew or could reliably guess at the plain text that corresponded to several letters of an intercepted Enigma message, a unique pattern was formed in the way those enciphered and unenciphered letters were interrelated from one position to the next in the message; moreover, those patterns had distinct mathematical properties which were unique—or at least nearly unique—for every different setting of the Enigma, and these could be tested systematically to find the setting that had generated them. The bombes used an electric motor to drive the rotors of a battery of interconnected Enigmas through every possible position until the looked-for pattern was electrically detected. Each “hit” could then be tested by setting up an Enigma replica with the setting recovered by the bombe and seeing if it worked to decode the rest of the message into something that looked like intelligible German.
The problem with the naval Enigma was twofold. First was getting enough reliable “cribs,” those bits of known contents of a message. It was a chicken-and-egg dilemma. Military wireless traffic usually contained an abundance of stereotyped phrases, such as times and dates of transmission, “from” and “to” lines, common abbreviations such as BdU. The process of matching a crib to its corresponding piece of cipher text was facilitated by a peculiarity of the Enigma machine, that a letter was never enciphered by itself: an A in the plain text never became an A in the enciphered message. So a possible crib could be slid along the coded message until a place was found where there were no “crashes,” spots where the same letter appeared at the same position in both plain and cipher text. But without being able to read a few messages to begin with, it was hard to guess what a good crib might be. Moreover, as would later become apparent, the German navy ran a very tight ship when it came to cipher security; wireless operators were instructed to vary common abbreviations and standard phrases, so that BdU for example might be written BDUUU or BEF.UNTERSEEBOOTE or BEFHBR.UUUBTE, or some other variation each time it was used. Finding a way into the system depended on getting a few successes to provide an initial foothold.37
The other problem was the navy’s indicator encryption system, which was the true nightmare. Breaking one army or air force message each day with the bombe was all that was needed to read all the rest of the day’s traffic on the same network. The basic setup of the machine changed only once a day, following a printed list. The individual setting used for each message was then varied by choosing a different starting position of the Enigma’s three rotors. The rotors each had 26
different positions, labeled A through Z with the letters of the alphabet, so there were 26 × 26 × 26 different possible start positions, 17,576 in all. The starting position for any given message was specified by a three-letter indicator, such as GHW or QYZ, and the simple system the army and air force used for encrypting these indicators at the start of each message allowed the Bletchley code breakers to read them off as easily as their intended recipient, once they had recovered the basic setting of the day using the bombes. The navy’s external code tables and more complex system for enciphering the message indicators, however, meant that breaking one message got the code breakers no further than that one message. A bombe run took hours to crack a single message; tackling every message individually was impossible.
With the code tables from U-110 in hand the whole problem split wide open. A flood of cribs for future work was provided along with a regular trove of decoded orders to U-boats and sighting reports sent from the submarines. With traffic in hand for May and June, Turing was able to develop a way to reconstruct future months’ code tables even without having any more lifted copies. Just to make sure, the Admiralty cautiously agreed to one more “pinch,” and on June 25 another German weather trawler, the Lauenburg, was captured in the Norwegian Sea about 900 miles north of Scapa Flow with the July list of daily settings aboard. (When the captain of the British destroyer Tartar explained to his crew that he wanted them to fire on the trawler but not hit her, the chief gunnery mate replied, “Christ, that should be easy.”)38
The code breakers soon discovered that some hand cipher systems used by the German navy, mainly to communicate with small ships in the Baltic, frequently carried weather reports and mine warnings that were repeated verbatim in Enigma messages sent to larger ships and U-boats. This became another fertile source of cribs for breaking future traffic; on several occasions the British were able to speed the process by laying mines with the deliberate purpose of generating a German message. By July and August 1941 information from naval Enigma signals more than once was available in time to divert a convoy around a waiting wolf pack. Sinkings by U-boats in August dropped dramatically to 80,000 tons.
CHURCHILL’S ENTHUSIASM for scientific invention was arguably exceeded only by his enthusiasm for cloak-and-dagger intrigue; that, too, caused its share of disasters throughout the war. Still, it meant that the prime minister needed no selling on the importance of the Enigma decrypts. In an initial burst of utterly unrealistic excitement, he demanded to have a copy of every decoded Enigma message delivered to him daily, in a special dispatch box. Later he settled for receiving a selection of the most important messages, but still insisted on seeing the actual texts, not summaries. He constantly bombarded the chiefs of staff and theater commanders with cables and memoranda calling their attention to what he considered significant bits of the decoded messages.
Churchill also took an early opportunity to show his support for the code breakers. On September 6, 1941, the prime minister paid a surprise visit to Bletchley Park. Standing on a tree stump by the lake on the grounds of the park, he addressed the assembled staff. “You all look very … innocent,” he jokingly began.39
The work was becoming much more technical as well as much more of an assembly line than anything that GC&CS’s old-school donnish linguists had ever known, and the breakthrough on the naval Enigma now brought the clash of cultures to a head. Bletchley’s director, Alastair Denniston, was still struggling to run things the old way; he was conscientious but simply overwhelmed by the logistical problems of recruiting, hiring, and housing hundreds of new staff. He reacted touchily to complaints from the naval Enigma group throughout the summer of 1941 that they were desperately short of manpower to maintain the bombes, carry out IBM punch card operations that were an essential part of the job, test the bombe results on the replica Enigmas, and even get their results typed up.
Simply, Bletchley was unprepared for success. Having achieved the near-miraculous in cracking the naval Enigma problem, they were now unable to exploit it to the fullest. In August the head of GC&CS’s naval section, Frank Birch, wrote to Denniston pointing out that the shortage of typists alone was causing their output to queue up for days. Administrators were wasting an inordinate amount of time on recruiting, but it was proving almost impossible to get young women in particular to work at Bletchley given the low salaries, lack of recreational facilities, inadequate meals, and poor housing, notably the primitive plumbing that was standard in private billets in the area—“no baths at all and the W.C. at the bottom of the garden.”40 All this did was to elicit a testy dismissal from Denniston of Birch’s “somewhat destructive memorandum,” as he termed it. “What does Birch suggest,” Denniston continued, “that we should move to Harrogate or some such place.… There are worse places in the country where there is not even a cinema. There is certainly one good cinema in Bletchley.”41
Meanwhile Dilly Knox was furiously resisting any attempts at mass-producing the process of reading enemy messages, fighting a rearguard action from his embattled ivory tower to preserve what he thought was his sole right to keep control of his discoveries, rather than collaborating as part of a team. “As a scholar, for of all Bletchley I am by birth breeding education profession + general recognition almost the foremost scholar,” he crazily wrote Denniston, “to concede your monstrous theory of collecting material for others is impossible … had the inventor no right to the development and production of his discourses, we should still be in the Dark Ages.… There are occasions when disobedience is a primary duty.”42
Something had to be shaken up. In October several of the key Bletchley mathematicians working on the Enigma decided on an audacious move. For reasons that he said he could not remember, Stuart Milner-Barry, an international chess master who had been recruited to Bletchley in early 1940 and was now deputy head of Hut 6, which was responsible for army and air force Enigma, was chosen by the others to carry a message to the prime minister himself. It was Trafalgar Day, October 21, 1941, the 136th anniversary of Nelson’s victory in 1805 over the French fleet. “What I do recall,” Milner-Barry said, “is arriving at Euston Station, hailing a taxi, and with a sense of total incredulity (can this really be happening?) inviting the driver to take me to 10 Downing Street.”43
The taxi driver did not bat an eye, and arriving at the prime minister’s residence Milner-Barry boldly marched in and announced he had come from a secret war station and needed to see the prime minister immediately on a matter of national importance. He was, unsurprisingly, told that was impossible. Milner-Barry countered that he could not possibly leave the letter he was carrying with anyone but the prime minister himself, given its sensitive nature. Finally, Churchill’s principal private secretary, Brigadier George Harvie-Watt, appeared; Milner-Barry was able to establish that he wasn’t a raving lunatic by referring to the prime minister’s recent visit to Bletchley, which Harvie-Watt knew of; and the secretary promised to see that the letter would be handed directly to the prime minister.
The code breakers emphasized in their letter that crucial work was “being held up, and in some cases not being done at all” due to the manpower shortages; the recovery of the naval Enigma keys was being delayed at least twelve hours each day; promises made back in July that the work of testing the bombe solutions would be turned over to a contingent of Wrens—members of the Women’s Royal Naval Service—had come to nothing. The next day the code breakers’ letter landed on the desk of General H. L. Ismay, the prime minister’s chief military aide, with one of Churchill’s famous red ACTION THIS DAY labels affixed. Beneath the label Churchill had scrawled a succinct instruction: “Make sure they have all they want on extreme priority and report to me that this has been done.”
Within a few months Denniston was out; orders for more bombes were issued; the Ministry of Labour was ordered to meet with the head of the British Secret Service and arrange whatever manpower Bletchley needed; the army, navy, and air force were told to make additional servicemen and servicew
omen available. Bletchley’s staff would reach about 1,500 by the end of the year, and from the naval Enigma group in Hut 8 a steady stream of decrypts flowed to the Admiralty’s Operational Intelligence Centre, where U-boat positions were updated with pins on a huge wall chart.
The combined effect of Bletchley’s breakthrough in reading U-boat messages and the growing efficiency of British antisubmarine patrols as they began to implement the recommendations of Blackett’s team delivered a sharp check to Dönitz’s offensive in the second half of 1941: the “Happy Time” was over. Monthly sinkings by U-boats, which had been averaging 250,000 tons in the first half of the year, dwindled to half that figure as convoys were safely rerouted around the lurking wolf packs, while the number of U-boats sunk doubled.
Puzzling over his abrupt change of fortune, in particular the failure of his boats to intercept one expected convoy after another, Dönitz confided his frustrations and suspicions in his war diary entry of November 16, 1941. “Coincidence does not fall on the same side every time,” he wrote. Yet nothing really seemed to explain it. It was possible the British had a new method of radio direction finding that was precisely locating the U-boats’ positions at sea; there were rumors about new kinds of radar; there was always the possibility of spies or treason. All seemed unlikely. The one thing the admiral was certain of was that it was impossible for anyone to have broken the Enigma cipher, given the sheer number of mathematical permutations and the safeguards used to encrypt the setting changes. “This possibility is continuously checked by the Naval War Staff,” Dönitz noted, “and regarded as out of the question.”44
Blackett's War Page 22
