5

 

Battle of the Atlantic: Cryptologic Seesaw

 

 

When the Germans went to war in 1939, their navy was almost completely unready to take on the Royal Navy of Britain. German naval officers had considered an outbreak of hostilities with the English an unlikely prospect and had planned accordingly. Their aim was to give Germany a small, balanced fleet that could play a useful subsidiary role in a war against another continental power—France, say, or the USSR. The few state-of-the-art battleships they had designed were meant as much to be showpieces of German technology and expressions of German pride as to serve as meaningful war machines. As for submarines, which had been so lethal a force in the Great War, the German navy had failed to build even as many as permitted in postwar armaments agreements, and the ones they did commission were small vessels with limited range.

Almost completely unready. One important exception has already been noted: the Germans' cryptographic branch, B-Dienst, was reading a variety of the Royal Navy's codes. Britain's tradition-bound Admiralty had rejected Lord Louis Mountbatten's recommendation that it change over to code machines, as the RAF and army had done. Instead, it was clinging to the age-old practice of relying on codebooks whose listings of naval terms and their cipher equivalents had to be enciphered and deciphered by operators using manual methods. The naval officers believed that by employing superencipherment tables and changing them frequently, they could thwart cryptanalysis. It was a vain hope, especially when German divers recovered current codebooks from sunken British warships. By the summer of 1940, B-Dienst cryptanalysts were deciphering a high percentage of the traffic in the codes used by both the Royal Navy and the Merchant Navy. They were reading a total of some two thousand messages a month.

The German advantage came to the fore immediately after the war began. B-Dienst's decrypts more than made up for the U-boat fleet's limitations. The subs sent their first British ship to the bottom only days after war was declared, and between September 1939 and the end of the year they, together with other ships of the German navy, sank 150 merchant vessels, the passenger liner Athenia, the battleship Royal Oak and the aircraft carrier Courageous.

They had another advantage in the competence of the man commanding the Unterseeboot, or U-boat, command, Admiral Karl Donitz. Donitz was convinced that his U-boats could strangle Britain into submission, as they had almost done in the 1914-18 war. Germany's neglect of U-boat construction during the interwar years dismayed him; he needed at least three hundred boats—and preferably one thousand—to do the job. But he set to work skillfully with the fifty-six operational craft available to him at the war's outset, while thirteen boatyards began turning out newer and better designs of U-boats in a stepped-up building program.

To these positives should be added the German navy's development of the most advanced and secure code systems of all the Enigma users. As mentioned earlier, the navy's Enigma was equipped with three extra rotors—eight in all. Although only three were used at any one time until later in the war, to have eight available meant that they could be arranged in 336 different ways and not merely 60 as with the army and air force Enigmas. Further, the navy's method for indicating message settings was entirely different. Instead of having the Enigma encipher the keys—a practice that had betrayed other services' settings—the navy instructed code clerks to determine their settings using a two-letter "bigram" table that was, in effect, an extra code unto itself. The bigram substitutions made the system "operator-proof": no choosing girlfriends' initials or sequences copied from the keyboard. The navy's operators weren't likely to take those shortcuts anyway; they were much more disciplined in observing strict security measures than the code clerks handling Luftwaffe and army traffic.

Further, the German navy used its Enigmas for a bewildering variety of codes. The list of merely the principal codes compiled by British naval intelligence officer Patrick Beesly in his book Very Special Intelligence runs to thirteen, ranging from the Atlantic U-boats' Triton to the Bertok code used for communications between the naval attache in Tokyo and the navy high command.

Dilly Knox and his team at Bletchley Park were inclined to give up on the navy Enigma. They had concluded that its system was impregnable. 

Coupled with this spirit of cryptologic helplessness was the fact that the British had woefully neglected the resources needed to conduct antisubmarine warfare. Control of the Admiralty had been taken over by Great War gunnery officers who believed that the scant funds available to the navy in the 1930s should be spent on battleships rather than convoy escort vessels or reconnaissance aircraft.

At the war's beginning, consequently, it appeared that Donitz and his U-boat crews held the winning hand. They might well fulfill their promise to Hitler to turn Britain's protective moat into a garrote around the nation's neck.

This early air of pessimism did not apply to Alan Turing and his colleagues in Hut 8. He refused to be discouraged. As Stephen Budiansky has expressed it, "During those two months Turing had performed a feat that if anything surpassed Rejewski's. He had laid out the entire theoretical framework for tackling the Enigma in all its variations." Turing also recognized the importance of seizing any opportunity, or even creating opportunities, for the capture of naval Enigma materials. Nevertheless, except for small temporary successes, Donitz's system defied Turing for the first twenty months of the war.

While his first bombe—the one without Welchman's diagonal board—was still being built, Turing went at the naval Enigma using manual methods. His approach, as ever, depended on applying cribs. What cribs could he hope to use to get at the naval Enigma messages?

One he borrowed from the Poles. They had found that often a message was a continuation of another and that this second message began with FORT, "forward from," followed by the time of origin of the preceding message. When it became known from a German POW interrogation that the naval Enigma procedure was to spell out numbers, the crib was extended. A message originating at 2330 could be enciphered as WEEPY. So the time that a message was sent could lead to identifying the plaintext letters. This method of cribbing was consequently called FORTY WEEPY. Using it, Turing and his small team managed to break messages from five days of Enigma traffic during the preceding November, before a more complex system of plugboard connections had gone into effect.

To break subsequent messages from those days, Turing invented what became known as the EINS catalog. He had observed that the word eins, German for "one," occurred in about ninety percent of all messages. By compiling a catalog of all possible encipherments of EINS, he was able to search out equivalents in the ciphertexts—another small opening into securing the plaintexts.

BP's "Prof" got his first boost from captured materials in February 1940. A German U-boat, U-33, tried boldly to lay mines in the Firth of Clyde, on the west coast of Scotland. The U-boat was discovered, depth-charged and driven to the surface. Mindful of the need to protect his Enigma, the skipper passed out the code wheels and insisted that when his crew members jumped into the sea they were to make sure that the wheels sank. The scheme didn't quite work. When the British rescued most of the crewmen, three code wheels were found in the pocket of a nearly drowned sailor. Two of the three were of the new code wheels added to the naval Enigma, and from them Turing learned the wiring of all but one of the new wheels.

Another helpful capture came in April 1940 when a British destroyer off the coast of Norway disabled a German trawler posing as a Dutch fishing vessel. Before surrendering, the German crew threw overboard two bags containing confidential documents and cipher materials. A British crewman dived into the icy water to rescue one of the bags, and other papers were found aboard the trawler. Included were the Enigma keys for the four previous days. With this gift Turing and his team were able, in May, to read some of Donitz's April traffic and to advance their understanding of the naval Enigma. But they could still not decipher further messages.

Stimulated by the captures, however, Turing's inventive mind came up with another development that he could put to use when his first bombe arrived. This was a technique which, used on intercepts for which no cribs could be imagined, drastically reduced the number of bombe runs needed to determine the arrangement of the Enigma's rotors for a specific time period. It was an essential step, since to run through all 336 possible rotor orders would take as much as a week of full-time operation. His limiting method became known as Banburismus because the long sheets of paper used in it came from the town of Banbury. With Banburismus, Turing exploited a flaw in the naval Enigmas. The Germans thought they were adding to their system's security by having each rotor notch that caused its neighbor to turn located at a different letter of the alphabet. Thus, rotor one had its notch cut at the letter R; number two at F; number three at W; number four at K; and number five at A. The BP analysts created a meaningless but memorable mnemonic to keep the sequence in mind: "Royal Flags Wave Kings Above." Turing saw these disparate notch settings as a way to identify each rotor—if the Germans had cut their notches all at the same letter the Banburismus method wouldn't have worked. Later the Germans recognized the flaw. When they added three new rotors, all were notched identically.

The Banbury sheets were made of stiff white paper about ten inches from top to bottom but varying in width from two to five feet. On them were printed vertical rows of alphabets side by side. The narrower sheets accommodated shorter messages. The young women of Hut 8 took each intercept and punched holes in the alphabets for each of its letters. By sliding the sheets over each other above a dark table, analysts could spot the places where the same letters appeared in a pair of messages. By keeping score on the repeats, Turing and his colleagues could, by a series of logical—if formidable—deductions, apply the Royal Flags formula to determine the lettered notches at which the rotors in use that day were located. With luck, Banburismus could reduce the number of rotor orders to be tested by the bombes to as little as six runs. A detailed explanation of Banburismus is included as an appendix in Sebag-Montefiore's Enigma book.

 

 

Ascendance of the Wolf Packs

 

For all of Turing's brilliance, Hut 8's breaks in the naval Enigma remained, throughout 1940 and into 1941, too few and too slow to be of tactical use against Dönitz's U-boats. It was a time of havoc for British shipping. Merchant ship losses for 1940 totaled nearly four million tons, when only one million tons of new shipping was under construction in British shipyards. From B-Dienst, Dönitz knew the convoys' routes and sent his U-boats out in "wolf packs" to waylay them. To make matters worse for Britain, with the fall of France in mid-1940 both the Luftwaffe and the U-boats gained the great advantage of relocating in western France. The German air force added its own substantial totals to drowned tonnages of Allied shipping. For the U-boats, the new French ports markedly increased the times they could spend prowling the North Atlantic. They enjoyed what their crews thought of as a "happy time." The anguish in Hut 8 was deepened by press reports of victorious U-boat crews returning to their Bay of Biscay bases to the blare of brass bands and the popping of champagne corks, not to mention the presence of the fiihrer himself, there to pass out Iron Crosses.

So desperate for new captures of naval Enigma keys did Turing and his team become that they supported a wild scheme put forward in December 1940 by Ian Fleming, a naval intelligence officer later to become the creator of James Bond. Fleming's idea was to seize one of the Enigma-equipped rescue boats the Germans manned in the Channel to pick up downed Luftwaffe fliers. To do this, Fleming advocated, and gained permission to implement, dressing an English flight crew in German uniforms and putting them aboard a captured German bomber. After the next raid on London, when the bombers headed home, the bogus bomber would sneak in among them. Over the Channel it would begin to emit fake smoke from its tail, send out an SOS and ditch. The crew would float in a rubber dinghy until deliverance appeared. "Once aboard rescue boat," his plan continued, "shoot German crew, dump overboard, bring boat back to English port." For the "word-perfect German speaker" among the English fliers, Fleming nominated himself.

The scheme almost came off. Fleming secured a plane from the Air Ministry, recruited "a tough crew" and settled in Dover to await his chance. The stumbling block was that neither reconnaissance flights nor radio monitoring turned up a suitable boat to be raided. The mission had to be aborted, much to the disappointment of all concerned.

BP did hatch an ingenious and successful stratagem for securing usable cribs. The codebreakers had discovered an interesting discrepancy in the Germans' use of naval codes when they dealt with mines sown by the British. German minesweepers were not equipped with Enigmas but relied instead on a manual code called the Dockyard Key, which BP had broken. When the British created new minefields, German warships had to be alerted, and they were equipped to handle Enigma traffic, not Dockyard transmissions. Oftentimes, when the RAF dropped mines, exactly the same message would be sent out in both ciphers. The new Bletchley scheme, appropriately called "gardening," was to have the Air Ministry sow a new minefield so that BP could reap the sudden sprouting of Dockyard and Enigma alarm messages. From the Dockyard decrypts, Hut 8 derived substantial cribs to be used in unlocking the naval Enigma. These cribs from two identical messages transmitted on different cryptographic systems were called "kisses."

In March 1941, Turing's team received another gift of captured materials. Once again off Norway, a trawler was disabled and driven ashore. This time the boarding crew seized the Enigma tables for the whole month of February. The resulting trove of decrypts was much larger. Moreover, Turing and his Hut 8 associates were learning enough about the naval Enigma to keep on reading the navy signals through the month of May. Then a new set of keys stymied them, and the flow of decrypts once more dried up.

None of this intelligence was of much use to the Admiralty. Messages that revealed important information were deciphered only long after it could have helped. However, the new batch of decrypts did yield one disclosure of supreme value. Young Harry Hinsley, studying the traffic, observed that some of the messages had been sent to small ships the Germans had stationed out in the North Atlantic to transmit weather reports back to headquarters. Although the weather trawlers were not enciphering their reports on Enigmas, they had to have the machines if they were to decode the Enigma messages transmitted to them. And since each ship was out there for stints of several months, they must have aboard a succession of naval Enigma tables.

Certainly, Hinsley told the Admiralty, if one of the weather ships was attacked, the crew would quickly jettison their Enigma and the current key tables. But wasn't it probable that future keys would be stored in a safe? And wasn't it also likely that in the stress of being attacked, the crew would forget about those hidden-away tables and they could be recovered by a boarding party?

The Royal Navy cooperated. A carefully planned foray on May 7, 1941, carried out by a flotilla of seven cruisers and destroyers, captured the trawler München east of Iceland. Sure enough, the weather boat crew, before surrendering, placed their Enigma and secret tables in a bag and threw it overboard. But also, as Hinsley had predicted, in the captain's quarters British sailors found the Enigma settings for the month of June.

Two days later, Bletchley analysts were handed an unexpected gift. A North Atlantic convoy escort group under Commander Joe Baker-Cresswell forced the German submarine U-110 to the surface. The U-boat skipper, certain that the sub was sinking, ordered his crew to abandon ship. But she didn't sink. Baker-Cresswell saw his opportunity. Remembering from his naval training how in World War I the capture of a German warship had enabled the British to seize codebooks and break the German navy's codes, he launched a similar mission aboard the U-110. He had the sub's survivors picked up and stowed out of sight belowdecks before his own crewmen boarded the U-boat and removed two boatloads of useful papers as well as the sub's Enigma.

The documents retrieved from the München and the U-110 gave Hut 8 a heartening, if temporary, breakthrough. Before, analysts had taken days to obtain Enigma decrypts. Now, with the June keys in hand, they lowered their decryption time to an average of six hours. The information they passed on to the sub trackers was current enough to locate U-boat wolf packs and deflect convoys to safer routes.

However, this heady time soon ended. The bigram tables they had so painstakingly compiled were replaced in mid-June, after BP had been reading Enigma traffic for only two weeks. Turing figured he needed a month of decrypts to work out the new tables. Could Hinsley persuade the navy to capture another weather ship?

He could. The navy would cooperate in tracking down the Lauenberg, which Hinsley selected as the most promising target. She had left her Norwegian base at the end of May to take over the North Atlantic patrol from a sister ship. Hinsley figured that she would be out there for at least a six week tour of duty and would consequently be carrying the Enigma keys for both June and July.

This time when the Royal Navy task force steamed off to find the weather ship, GC&CS staffer Allon Bacon went along. The operation came off smoothly. Informed about the little trawler's location by BP decrypts, the British flotilla found her behind an iceberg. The British fired shells that came close but deliberately avoided hitting the trawler, forced her crew to take to lifeboats, took the Germans aboard and sequestered them so that they could not see what the boarding party was doing. The Germans had ditched their Enigma and tried, only partially successfully, to burn the confidential papers in the ship's stove. The boarders presented Bacon with thirteen mail sacks of documents. Sorting through them, he found just what he hoped for: the Enigma settings till the end of July.

Once again, Hut 8 needed only a few hours in which to decipher the German navy's Enigma traffic. Of much larger significance, the learning curve had been completed. Turing's team could now proceed on its own, without need for further captures.

An additional benefit came from the U-110 materials. Besides the regular German naval Enigma, Donitz used an "Offizier," or "Officer," code for higher-level communications. It was doubly enciphered, first in the special code and then again in the regular Enigma cipher. The U-110 papers included the special Offizier settings for the month of June 1941 and so allowed Offizier messages to be broken for that month. But when the settings were replaced, BP was again shut out—that is, until Rolf Noskwith, a German Jew who had emigrated to England with his parents in 1932, studied the earlier decrypts and hit upon a crib that he used to crack open the code. Offizier was broken for the remainder of the war.

For the last half of 1941, and until February '42, the cryptologic battle of the Atlantic swung sharply in favor of the British. While B-Dienst was still breaking the Admiralty's codes, Bletchley Park had an ally whose contributions more than made up the difference. This inadvertent aide was Admiral Donitz himself. Early in the war the German admiral had decided, as he wrote later, that "I myself could quite easily direct the whole tactical operation against a convoy from my headquarters ashore." The result was a flood of communications back and forth as he instructed the U-boats where he wanted them to be and what he wanted them to do, while also expecting that they transmit regular situation reports back to him. The masses of advice BP was teleprinting to the Submarine Tracking Room far exceeded in value whatever B-Dienst was disclosing.

The statistics of shipping losses reflected the change in cryptographic fortunes. From April to June 1941, another 150 Allied ships were sunk. The July-to-September period showed a decrease to ninety, and from October to December the score declined to seventy. Factors other than the code-breaking helped account for the decline. The support by convoy escort vessels was strengthened, especially by the deployment of fifty World War I destroyers that the U.S. had taken out of mothballs and turned over to the British in a nominal exchange for ninety-nine-year leases on British bases in the Americas. Longer-range planes gave the convoys air cover for larger stretches of their journeys. Technological developments such as shipborne radar improved the convoys' chances. But the codebreakers had the most impact on the change. Their decrypts enabled the convoys to be rerouted, often in midcourse, to swing clear of the wolf packs. By pinpointing the U-boats' locations, BP decrypts caused more of them to be sunk. From a seemingly hopeless struggle, the Battle of the Atlantic, for this sunny period, swung in favor of the Allies.

 

 

BP Decrypts Help Sink Surface Raiders

 

While the U-boats posed the most serious threat to Allied shipping, they were by no means alone. The marvel warships the Germans had begun building in the 1930s stood ready to be released as hunters and raiders. The Sigint forces at Bletchley Park paid special attention to them because they were new and fast and, therefore, especially dangerous.

The first major engagement with the raiders came too soon for BP to have much effect. In December 1939 the pocket battleship Admiral Graf Spee was set loose in the South Atlantic, sinking three British merchant ships in five days. She was tracked down by three Royal Navy cruisers, whose guns scored some fifty hits, forcing the raider to take refuge in the harbor of Montevideo, Uruguay. Seeing no possibility of escape, the ship's captain ordered her to be scuttled. He then shot himself.

In the aftermath, though, British intelligence did score a notable success. Accompanying the Graf Spee was a supply ship, the Altmark. Before scuttling the battleship, the German captain ordered that some three hundred captured British merchant seamen be put aboard the Altmark, which was to be dispatched to Europe. British and French agents tracked her to Norway, where a Royal Navy destroyer intercepted her and liberated her prisoners before sinking her.

Admiral Erich Raeder, the German navy's commander in chief, next planned for his surface raiders to attack Atlantic shipping. He would unleash his newest battleship, the Bismarck, accompanied by the heavy cruiser Prinz Eugen. They would be joined by the battle cruisers Scharnhorst and Gneisenau. His plans were partially frustrated when Scharnhorst needed to stay behind for engine repairs and Gneisenau was severely damaged when a gallant Coastal Command torpedo plane pilot got off his missiles before being shot down. These setbacks did not deter Raeder from sending out the Bismarck and Prinz Eugen.

Since naval Enigma messages were then being broken only days and weeks after being intercepted, they were usually too late to be of use. However, an April message read in May did disclose that the Bismarck was to be loosed as a raider. Aerial recon found that she had first moved to the Norwegian port at Bergen and, subsequently, that she had left Bergen.

A British naval squadron located the Bismarck and engaged her, but she shot her way out, sinking the battleship Hood and damaging the Prince of Wales. She herself suffered a bomb blast that caused her to leak oil. In foul weather, the Royal Navy lost contact. When the German warships separated, British attention focused on the Bismarck. Where would she be headed? After escaping past the north of Britain toward Iceland, would she double back to Norway? Or keep on going in the Atlantic? The British could only guess.

Harry Hinsley weighed in with his opinion. He lacked decrypts for an answer, but he could see from direction-finding reports that messages being sent to the Bismarck, previously originating in Wilhelmshaven, were now being sent from Paris—a sure sign, to him, that the ship was bound for a French haven.

Admiralty authorities shook him off. His information was too tenuous to be relied on. August minds at the top were convinced that a more northerly route was most likely.

BP provided the clincher. Normally Gordon Welchman's breaking of the Luftwaffe's Red code would not have supplied any useful naval information, but this time it so happened that a Luftwaffe general in Athens had a close relative on the Bismarck, and he inquired as to the ship's destination. The word came back and was promptly decoded in Hut 6: the Bismarck was bound in a southeasterly direction for the French port of Brest.

She never made it. Even though one British flotilla misinterpreted the signals that intelligence transmitted to it and headed off in the wrong direction, the battleship Ark Royal, hurrying up from Gibraltar, found her. One of its aircraft torpedoed her, jamming her rudder so she could only travel in circles. Her captain ordered her to be scuttled.

As with the Graf Spee, supply ships had been dispatched to the Atlantic, both to serve the needs of the Bismarck and to aid the U-boats. By early June, Hut 8 had begun reading naval Enigma messages almost currently, and decrypts revealed the positions of eight supply ships. To avoid arousing the Germans' suspicions that the codes were being broken, it was decided that only six of the eight ships would be destroyed. As luck would have it, though, other Royal Navy ships, without advice from BP, happened upon the remaining two and sank them as well. Concern about Ultra was soon eased. The Germans continued to blame the sinkings on superior British direction finding or French agents or some other cause rather than accept that the Enigma could be compromised.

The Scharnhorst and the Tirpitz, the last two of Germany's battle cruisers, were hiding out in Norwegian harbors, ready to sail out and ravage convoys headed for the USSR. At Christmastime in 1943 the Germans decided to unleash Scharnhorst for an attack on convoy JW55B. Donitz, succeeding Raeder as commander in chief, was sensitive to the plight of German soldiers now being battered on the Russian front and wished desperately to stanch the flow of supplies to the Soviets. He believed he could commit the precious battleship because he had lulled Britain into more relaxed convoy protection by allowing recent sailings to pass through unmolested. Bletchley decrypts, however, warned of his plans. Although key German messages relating to the Scharnhorst were in the Offizier code, whose doubled encipherment slowed decryption, and were of little tactical value, BP did inform the escort fleet when the Scharnhorst put to sea. Further, the decrypts guided the British warships in placing themselves to intercept the German battleship when she did try to attack the convoy. In the early hours of December 26, a star shell fired from a British cruiser lit up the Scharnhorst instants before a salvo from the fleet's battleship, HMS Duke of York, slammed into her. Left burning, she was finished off by torpedoes from swarming destroyers. Of her crew of more than two thousand, only thirty-six men were rescued from the Arctic waters.

The final chapter of warfare against Germany's surface raiders was not written until November 1944. This was the struggle against the Tirpitz. Fear of this capital ship had caused the British to attempt repeatedly to put her out of commission. In March 1941 a commando raid against Saint-Nazaire succeeded in seriously damaging the only dry dock on the Atlantic coast capable of handling repairs of the massive battleship. In October 1942 the Admiralty sent across the North Sea a Norwegian naval officer in a fishing boat that had two manned torpedoes slung beneath it. The idea was to enter the fjord where the Tirpitz was berthed and aim the torpedoes at the ship. The scheme went awry when the boat encountered a squall and lost both the torpedoes. In September 1943 a fleet of six midget submarines was towed into position to glide in close to the Tirpitz and her escort vessels and detonate explosives beneath their keels. Three of the tiny subs made it through and, as Enigma decrypts later confirmed, made hits on the Tirpitz, so badly damaging her that the Germans did not expect to have her ready for action until mid-March of 1944.

When, repaired, the ship again became a danger, she was attacked and once more damaged, this time by carrier aircraft. A second carrier-planned sortie in July, meant to finish her off, met with increased antiaircraft fire and made no hits. But on September 15, 1944, RAF bombers operating from north Russia again put her out of action.

With Ultra decrypts carefully monitoring every stage of the big ship's latest repairs, the finale came when Tirpitz was once more reported ready for raiding. On the twelfth of November, thirty-two British bombers, each carrying a single twelve-thousand-pound bomb, took off from Scotland. At least two of the bombs hit the Tirpitz and capsized her. German sailors, hopelessly trapped within the inverted hull, were heard singing the German national anthem, "Deutschland Uber Alles," to their last breaths. "What a tragedy," one observer commented, "that men like that had to serve the Nazi cause."

Removal of this last threat of the surface raiders enabled the Admiralty to dispatch warships to the Pacific for the war against Japan.

 

 

Bletchley Copes with Shark

 

In February 1942, when everything had been going so well, Hut 8's intelligence feast abruptly ended. Up until then, both the surface ships and the U-boats had used a common cipher, called Dolphin by GC&CS. Now the U-boat command gained its own cipher, which the British dubbed Shark. It was not only a new code; it involved a change in the design of the U-boats' Enigma machines. This was a thinned reflector that allowed a fourth rotor to be added in its slot. Applied to the machines Donitz used in his communications with his Atlantic and Mediterranean U-boats, Shark plunged Bletchley into a ten-month blackout. During those long months, Hut 8 penetrated the new cipher on only three days—and each of these times only because of a German error in sending a message in both the Dolphin and Shark ciphers. Banburismus no longer worked. Turing's EINS catalogs were of no avail. The Submarine Tracking Room reported, "Little can be said with any confidence in estimating the present and future movement of the U-boats."

What made those ten months doubly frustrating was that by breaking codes other than Shark, BP knew a great deal about the U-boats: the commissioning of new craft and their trial runs in the Baltic, their expected performance, their armament, the experience levels of their commanders, their transfers to the west of France, even the times of their departures for active duty and their arrivals back in port. What was lacking was the most important information, in Hinsley's words, what happened "between the time they left harbour and the time they returned from patrol."

To make this long blackout period still more disastrous, B-Dienst was again gaining the upper hand in the contest of which antagonist was breaking the other's codes. The British Admiralty had switched to a new nonmachine code, but B-Dienst was readily reconstructing it. Donitz knew the schedules of the North Atlantic convoys and the courses they would take.

The Sigint seesaw had peremptorily swung back to the German side—at a time when Reich production was delivering increasing numbers of U-boats into Donitz's hands. In addition, he was receiving large "milk cow" supply submarines, each of which could deliver seven hundred tons of spare fuel and torpedoes, saving the U-boats the forty-six-hundred-mile round-trip back to their bases.

The consequences were muted for a while by the fierce winter of 1941-42, for its violent seas decreased the U-boats' effectiveness. When the weather moderated, however, Donitz and his commanders made up for lost time. They steadily increased their sinkings, while their losses of U-boats declined. The first half of 1942 resulted, again in Hinsley's words, in the U-boats' "greatest sustained period of success in the whole course of the war."

In view of the sudden inability of the convoys to steer clear of wolf packs coinciding with a change in code, the U-boat command might well have been tipped off that their earlier codes had been broken, if not for one momentous change in the war's course: the entry of the United States into the conflict. Four days after Pearl Harbor, Hitler had honored his Tripartite pledge to the Japanese by declaring war on the U.S. All that American shipping he had placed off-limits to the U-boats in order not to provoke the U.S. now became fair game.

As a result, instead of continuing to send his boats against the Atlantic convoys and finding them suspiciously vulnerable, Donitz directed his boats against the coasts of the U.S. and Canada. He did not have as many to dispatch as he would have liked. Many had to be diverted to the Mediterranean to help protect the supply lines to Rommel in North Africa. In addition, Hitler, still fearing a flank attack through Scandinavia, had ordered other U-boats to patrols in Norwegian waters. For attacks on U.S. shipping, the German admiral had only a "handful of boats, but they were the large, long-range Type IX craft now being supplied by German boatyards.

American unpreparedness enabled him to make the most of what he had. The U.S. could assemble few vessels suitable for escort duty, and the ones available were not assigned to convoy protection. The navy's head, Admiral Ernest King, did not approve of convoys; he preferred to go after the U-boats in open-sea hunts. These forays, however, proved fruitless. The U-boats simply hid out on the bottom until King's patrols had passed before rising to continue their slaughter of the busy traffic along the eastern coast. U.S. coastal cities, resisting the inconvenience and possible trade loss that would ensue from blackouts, kept their lights undimmed, providing the German prowlers neatly silhouetted targets. In just two weeks the U-boats sank twenty-five ships totaling more than two hundred thousand tons, a high percentage of them tankers, and continued the sinkings at roughly one a day. Crowds of watchers along the coasts witnessed the deadly pyrotechnics of exploding ships. Another "happy time" for the U-boats had begun.

If Dönitz's approach to sea warfare had a fault, it was in his belief that what mattered most was the tonnage sunk by his subs. Despite knowing that choking off British supplies in the North Atlantic was the real key to German victory, he was unable to resist the opportunities to pile up tonnage records elsewhere. His impressive statistics, it must be remembered, made for status-saving, job-preserving reports to Hitler. Yet while his U-boats were scoring easy points in U.S. waters, massive convoys were passing through to Britain almost unmolested.

Not until the summer of 1942 did the situation begin to change. By then Admiral King had given up the hunt missions and agreed to convoys. Coastal cities were blacked out. Plus, the navy established what it called its "bucket brigade." Tankers and merchant ships traveled up the coast in protected convoys by day and holed up in sheltered ports at night. The British helped by sending over escort corvettes and a squadron of RAF Coastal Command planes. Guided by visiting Britons, the U.S. Navy had begun setting up a Submarine Tracking Room similar to that of the Royal Navy. The happy time came to an end.

Dönitz rerouted his wolf packs to the North Atlantic. With few U-boat losses and strong inflows of new boats, his fleet had grown four times as large as when Shark had been introduced. Also he had found a chink in the Allies' defensive armor. This was the "Air Gap," a distance of three hundred miles between the extremity of air cover from Newfoundland and Iceland and that extending from the British Isles. In this gap he formed his boats into "picket lines." Aided by B-Dienst's decrypts, the pickets could detect approaching convoys and alert other subs to swarm in for the kill. Before Shark, when BP was breaking the naval Enigma, only one in ten convoys was sighted by the wolf packs. Now with BP blind, they found one of every three.

The carnage in the North Atlantic marked the second powerful German surge toward victory. In the first of the new round of convoy battles, eleven out of thirty-three ships went down. In the two months of September and October 1942, forty-three ships were sunk. By November the losses soared to 743,321 tons, the highest figure for any month in the entire war. During 1942 more than eight thousand merchant sailors were killed. Two of the ships sunk were carrying U.S. servicemen to England, adding to the lives lost.

The same grim story held true for convoys trying to deliver armaments and supplies to the Soviet Union. Grimmest of all was the fate of convoy PQ17, which set out from Iceland on June 27, 1942. The convoy was attacked in the Barents Sea by U-boats and aircraft. At that time the Tirpitz was still available as a raider. The convoy was given the misguided order to scatter. It didn't matter that the Tirpitz never got into the action. The U-boats and planes picked off the dispersed merchant ships one by one. Of the thirty-seven ships in the convoy, only thirteen reached Russian ports. As a result of the disaster, all convoys to Russia were suspended during the spring and summer of 1943.

The mounting destruction by the U-boats cast a pall of despair over news of the war that was otherwise turning in the Allies' favor. The British Eighth Army had defeated Rommel at El Alamein. The Allied landings in northwest Africa had surprised the Germans. The Germans' decision to occupy the whole of France had prompted the Vichy government to scuttle the French fleet at Toulon. But without Ultra's help, the situation in the North Atlantic was threatening to undo all the other triumphs.

Breaking Shark, Turing saw, required four-wheel bombes. Tabulating Machine Company engineers worked on their development, introducing some limited use of electronic tubes to speed their operation and race through the increased number of permutations introduced by the fourth rotor. Until these bombes could be delivered, Hut 8 could only wait. Luckily, an even harsher winter than the one preceding it slowed the U-boats in January and February 1943.

In March, however, they stormed back in force. During just the first twenty days, ninety-seven ships were sunk, with over half a million tons of supplies sent to the bottom. The official Admiralty verdict was that "the Germans never came so near to disrupting communications between the New World and the Old" as in those twenty days.

Behind the scenes, though, important changes were occurring. For one, the sheer productive might of the U.S. was tipping the scales. The Ships for Victory program was turning out standardized Liberty ships at a rate of three per day, enabling the Allies to produce more vessels than the Germans were sinking. The American merchant craft were faster, ensuring that convoys could move more swiftly. Escort patrol ships, aircraft carriers, destroyers and larger warships were being delivered by American shipyards. Recon aircraft with ever greater ranges were issuing from American factories in unprecedented numbers. Long-range American B-24 bombers were joining with British bombers to extend air cover over the convoys and close the Air Gap.

In addition, convoys were benefiting from British and American technology. Radiotelephones were installed to improve communications between ships and to coordinate their maneuvers. Escort vessels were equipped with their own direction-finding equipment to help them home in on lurking U-boats. Some larger freighters were fitted with airplane catapults, from which game pilots took off knowing they would, after their search-and-destroy missions, have to reach a land base or ditch near an Allied vessel in the hope of being picked up. Airborne radar and powerful new searchlights enabled Allied planes to detect and swoop in on U-boats traveling on the surface at night.

On December 13, 1942, the decisive change came. Shark was finally broken. This resulted from one last, all-important capture of German code materials and from the clever use Bletchley made of them.

The capture occurred in the Mediterranean, off the coast of Egypt. A British flying boat's crew sighted U-boat U-559 and alerted four destroyers to pursue it. Later that night, a depth charge forced the sub to the surface, virtually under the guns of the HMS Petard. The U-559's commander ordered his crew to abandon ship. Before escaping, the U-boat's engineer opened the sea cocks to scuttle the sub. But she remained afloat, her conning tower just visible above the waves. Four of the Petard's, crewmen either swam to the sub or jumped onto it from the deck of their ship—survivors' accounts vary. Three of the crew, Lieutenant Tony Fasson along with Colin Grazier and Ken Lacroix, clambered into the sub's interior while the fourth, young canteen assistant Tommy Brown, ran up and down the conning tower ladder in order to hand over to a whaleboat what the others could deliver to him. They grabbed the four-rotor Enigma from the radio room and an armful of charts and papers that Brown managed to transfer to the whaleboat alongside. At that moment the U-boat went under. Lacroix just managed to escape up the conning tower; Fasson and Grazier never made it.

Their sacrifice gave the Hut 8 team what it needed to crack Shark. In adding the fourth rotor, the Germans had taken into account that at times, in order to communicate with three-rotor machines, that rotor would have to be put in a neutral position—as, for example, when the U-boat had to communicate with a shore weather station. Among the papers the brave men of the Petard had delivered were the current editions of the three-rotor codebook for the Short Weather Cipher and the four-rotor U-boat key. The result was that when Shark was used for weather signals, the three-rotor bombes could be used to decipher the messages, and the remaining part of the day's key could be reconstructed by testing no more than twenty-six letters of the nonrotating fourth rotor. In the first hour after this breakthrough, a message revealed the positions of fifteen U-boats.

Those in Hut 8 felt both deep relief and huge elation. Pat Bing, then a teenaged typist, later recalled the excitement of finally being able to tap out German text on long strips of sticky tape, fasten the tapes to paper and send them by the compressed-air tubes the young women called "Spit and Suck" to Hut 4 for the interpreters and disseminators to work on. The deciphered Shark messages were, she said, "a great gift from the brainy boys' department."

Historian Patrick Beesly has recorded the impact of the conquest of Shark. The flood of decrypts and translated signals that poured into the Admiralty's Operational Intelligence Center, he noted, "made it possible, for the first time, for . . . the Submarine Tracking Room to build up a comprehensive and accurate picture of the whole operational U-boat fleet."

At this point the cryptologic war reached a stalemate. BP was reading Donitz's copious exchanges with his U-boats, but B-Dienst was reading the Admiralty's output. As Hinsley put it, "Between February and June 1943 the battle of the Atlantic hinged to no small extent on the changing fortunes of a continuing trial of cryptographic and cryptanalytic resourcefulness between the B-Dienst and the Allies."

There was one significant difference. From their decrypts, Bletchley's cryptanalysts gained unmistakable proof that the Germans were breaking the Royal Navy's main code, and so set in motion the changes necessary to provide a more secure system.

It took the Admiralty until June to make the changeover. In the interim, U-boat warfare rose to its savage climax. As an example, during four days in mid-March, Donitz's wolf packs hurled themselves against two intermingled convoys, HX229 and SCI22, whose course had been plotted by B-Dienst, and sank thirty-two of their ships plus a destroyer, with a loss of only one U-boat. It was the greatest U-boat success of the war. Again the Admiralty was reduced to despair, even to considering that "we should not be able to continue convoy as an effective system of defence."

April produced a standoff. Shark decrypts enabled the Admiralty to reroute threatened convoys, but B-Dienst decrypts informed Donitz how to counter the instructions and reposition his boats to the best advantage. So exhausted were his crews and their equipment by the March onslaught, however, that he could not maintain their previous level of sinkings. April's toll dropped to 277,000 tons.

Then came what those in U-boat command regarded as "Black May." Two calamities struck the U-boats. One was a sharp rise in their own losses: thirty-one boats were sunk during the month, and the total for the first six months of 1943 rose past one hundred. The second was a wavering in morale. Less experienced commanders exhibited a drop-off in zeal and a rise in caution compared with their predecessors. Donitz was driven to increasingly shrill denunciations of his crews for their failures to press home their attacks.

The turn of the tide was dramatized by the passage of convoy SCI30 in mid-May. Though attacked by a pack of U-boats, not a ship was sunk. By contrast, six U-boats were lost and others damaged.

The price was more than the German admiral, who had lost his own son in one of the downed boats, was willing to pay. On May 24 he sent out orders for his U-boats to withdraw from the North Atlantic and shift to less hazardous patrols southeast of the Azores.

Dönitz refused to concede, however. He later wrote, "Wolfpack operations against convoys in the North Atlantic . . . could only be resumed if we succeeded in radically increasing the fighting power of the U-boats." He was determined that in the autumn of 1944 he would launch a new campaign, using U-boats equipped with new developments from German science. The improvements included superior radar, better antiaircraft protection and more efficient acoustic torpedoes.

It was an abortive effort. By then the odds were stacked overwhelmingly against Dönitz. Long-range aircraft made the air cover for convoys complete, especially after Portugal permitted the Allies to occupy the Azores. Escort carriers were plentiful. New technological developments included "hedgehogs" that enabled destroyers to throw depth charges ahead of their course as well as behind. To counter the acoustic torpedoes, the Allies perfected the "Foxer," a device towed astern of the escort vessels; it attracted the torpedoes and caused them to explode harmlessly.

Above all, the cryptographic advantage had swung completely to the Allies. As Beesly expressed it, Dönitz "was now groping in the dark while our picture was so clear that convoys could be converted, Support Groups transferred, air cover increased or reduced in accordance with the daily or even hourly demands of a situation."

Few Allied ships were sunk, and too many U-boats were destroyed. On November 16 Dönitz ordered another withdrawal.

In the meantime, control of Shark had passed to American cryptanalysts. With a flood of super-high-speed bombes being produced by the National Cash Register Company in Dayton, Ohio, the Yanks were better equipped to deal in a timely fashion with the vast numbers of permutations that had to be run through to reach Shark Enigmas' settings. During the second half of 1943, National Cash delivered seventy-five bombes, more than the British produced during the rest of the war. American analysts informed the Tracking Rooms in London and Washington.

Dönitz still would not give in. He pressed U-boat designers to apply new techniques in a series of Super U-boats. Informed by Shark decrypts, British and American commanders watched these new developments with grave misgivings. The new boats' streamlined hulls and quiet new electric motors allowed them to slip along underwater at speeds matching those of most Allied escort vessels. Most worrisome of all, they were equipped with snorkel devices that took in oxygen and recharged batteries as the boats traveled at hard-to-spot periscope depths, enabling the subs to stay submerged for up to ten days.

Dönitz planned his Super U-boat convoy battles for early 1945, but his ambitious plans were thwarted as Allied bombardments of U-boat assembly plants and bases caused delays. Then the finished boats revealed flaws that had to be corrected. Before he could deploy his new subs, the war ended.

The Battle of the Atlantic was the longest sustained conflict of the war, one that cost both sides heavily. The Allies lost 2,603 merchant ships and 175 naval vessels. The lives lost exceeded 40,000, including 26,000 civilians. German U-boat losses numbered 784, killing 28,000 crewmen—two-thirds of the total force. The casualty rate was the highest suffered by any service during the war.

By June 1943, Churchill wrote, "The shipping losses fell to the lowest figure since the United States had entered the war. The convoys came through intact, and the Atlantic supply line was safe."

 

 

For Some Codebreakers, Unhappy Endings

 

Triumphant as they were during the war, both Alan Turing and Gordon Welchman suffered ill-fated deaths in the years that followed. Turing's fondness for Christopher Morcom turned out to be more than an adolescent aberration. He slowly surrendered to his true nature—at a time when British law still regarded homosexuality as a crime. Arrested after a petty encounter with the police, he was unable, because of his secrecy pledge, to assert his heroic wartime stature in his defense and was subjected to a humiliating trial and a judgment that his "cure" include the injection of female hormones that made this sturdy man who had run marathons become obese and grow breasts. Not long after, at the age of forty-two, he committed suicide.

Welchman, after the war, emigrated to the U.S., became an American citizen and served as a consultant on intelligence security during the Cold War. He became concerned that Allied codebreakers were making the same mistakes that had betrayed the Enigma. When Winterbotham's The Ultra Secret broke the walls of secrecy, Welchman felt released from security restrictions and had his book The Hut Six Story published by an American firm. But whereas Winterbotham had said little about the actual codebreaking methods used at Bletchley Park, Welchman, as a warning to current cryptographers, spelled out the "comedy of errors committed by the Germans" and how these errors had been exploited by Allied cryptanalysts.

Because of these disclosures, the book met with a storm of protests by both British and American authorities. The British banned publication in Britain and issued criminal charges against Welchman. The Americans withdrew his security clearance, making it impossible for him to continue his employment. Because of the harassment, Welchman wrote, "my health was seriously affected." British historian Nigel West has claimed that the persecution drove Welchman to a "premature death."

For both Turing and Welchman, however, these postwar troubles cannot dim the glory of their wartime achievements. The techniques they developed became the machinery of a huge intelligence factory at Bletchley Park, which ran so smoothly that in the war's latter stages they had only to oversee its almost routine production of war-winning decrypts by the hundreds of thousands.