by Kahn, David
The team was greeted at the end of its trip by the then No. 2 at B.P., Commander Travis, who took charge of the crates. The Americans were driven up to Bletchley, arriving at night. The big house was barely visible, with not a glimmer of light escaping through its blackout curtains. The men were led through the main doors, a blacked-out vestibule, and dimly lit hallways into Denniston’s office. He and his senior staff were standing in a semicircle around his desk. In a memorable moment that marked the renewal of a fruitful cooperation, the Americans were introduced to each person in turn. After some pleasantries, they were driven off to Shenley Park, the country estate of an oil magnate; the next morning, they were fed a magnificent English breakfast just as if there were no rationing.
For the next five weeks, the American team not only spent time at Bletchley but visited the direction-finding network, the Admiralty’s communications center, the O.I.C., the intercept stations at Scarborough and Flowerdown, and the radar at the underground command center beneath Dover Castle. At Bletchley, Rosen and Sinkov taught the British about PURPLE and took copious notes on the bombes and details of Enigma cryptanalysis. The British had few solutions to show, since up to the end of February Hut 8 had broken only eleven days of naval Enigma traffic, and the Krebs material arrived only five or six days before the Americans were to leave for home. But they did explain the attempts made to discover the machine settings, and they gave the Americans all the keys that had been recovered. And though the British never provided an Enigma, they did give Weeks a paper analogue. Sinkov and Rosen concluded in their official report: “We were invited to ask questions about anything we saw, no doors were closed to us and copies were furnished of any material which we considered of possible assistance to the United States.”
By the time of Pearl Harbor the U.S. Navy was putting 20 percent of its interception effort and 3 percent of its cryptanalytic effort into German and Italian naval systems (all the rest of the work was going into Japanese naval and diplomatic codes and ciphers). For a year after the four Americans left B.P. in the spring of 1941, no Britons visited the U.S. Navy cryptanalysts, and no Americans went to B.P., though the two had been exchanging direction-finding bearings since May of that year. In April 1942, with the United States in the war, a conference in Washington between the communications intelligence specialists of the two nations led to increasing collaboration in the form of telecommunications, visits, and permanent liaison personnel. Winn, who had spent a year at Yale and a year at Harvard after graduating from Cambridge, persuaded the Americans to set up a system like his. Then, in the fall of 1942, Lieutenant Joseph Eachus, U.S. Naval Reserve, was sent to England.
A tall midwesterner, Eachus, who held a doctorate in mathematics from the University of Illinois, had taken a correspondence course in cryptology from the navy. After Pearl Harbor, he was called up and, on his first day, was sent to a large room in the Navy Department building. It was bustling with activity. One man was practicing Morse code by sending to himself. One was practicing Japanese by talking to himself. Another was trying to call himself on another telephone in the room—via Alaska. Said Eachus, when all of this was explained to him, “Fm going to like it here!” This was OP-20-G, the Navy’s communications intelligence section: OP because it was a division of the office of the chief of naval operations, 20 the number of the communications division, and G for communications intelligence. When the section needed someone to go to England, Eachus could be spared, since he hadn’t been there long enough to have an important job.
For months Eachus was the only American at B.P. He studied the British cryptanalysis, reported on it to OP-20-G, and assisted in codebreaking. When he came home for a while in the spring of 1943, he found that OP-20-G had moved into new quarters in a former girls’ school, Mount Vernon Seminary, at 3801 Nebraska Avenue in northwest Washington. Here the cryptanalysts of OP-20-GY(A)—Y for the cryptanalysts, A for Atlantic—received intercepts from the posts scattered along the East Coast.
The codebreakers worked in what was called the “back room” but was actually three small rooms on the ground floor of what was called Building 2. One room was for the four women on each watch who sorted intercepts; both U-boat and U-boat Command messages could be intercepted on the East Coast. Another room housed the teletypewriters and cipher machines for communications with the intercept posts, G.C.&C.S., and the U.S. Navy’s U-boat plot. The cryptanalysts worked on the Enigma intercepts in the third room. Headed by reserve Commander Howard Engstrom, a Yale mathematician, they cooked up cribs from such stereotypes as the standardized German beginnings of messages—“It’s much easier to say the same thing the same way every day,” observed one cryptanalyst—and a weather report put out daily for the Bay of Biscay, which the British intercepted and sent to Washington.
The cryptanalysts had other sources of cribs as well. Once an intelligence officer sent a memo to Engstrom:
The Star of Suez was torpedoed and sunk 15 December at 01 N 29–30 W. The sinking was claimed in the German broadcast of December 17. We have no message relative to the sinking, which was in the area being patrolled by [U-134 skipper Lieutenant Rudolf] Schendel…. A DF [fix] at 0019/17 [December] (TOO [time of origin] 2301) of 63 groups, fixed at 03–45 N 29 W, is the most likely message that reported the sinking of the Star of Suez, though it might be a report on the sinking of the East Wales, at 2025/16 at 00-24 N 31-27 W, also claimed in the same German broadcast. Possibly the names of these ships, or routes Trinidad to Capetown and Trinidad to Durban, respectively, with the skipper’s name Schendel would afford a crib on the traffic for the 16th.
This crib may not have worked, for, as was later learned, a different U-boat had sunk both ships. But when one crib failed, the cryptanalysts tried different suppositions.
The codebreakers worked in close collaboration with Hut 8. The British forwarded messages that could not be intercepted in the United States, and the teams on both sides of the Atlantic sought cribs, dividing up the work by days. When the British found a crib, they transmitted the text of the intercept followed by the text of the crib; from this the Americans reconstructed the menu. If the Americans recovered a key, they would send it to Bletchley, where the British used it to decipher their intercepts. The two units communicated by radio and cable, the latter passing through the office of British Security Coordination in Rockefeller Center, New York. Secrecy was ensured by the CCM, the Combined Cipher Machine, which was not actually a single machine but, on the British side, the Type X cipher machine with an adapter that enabled it to decipher messages from the American machine, the ECM, or Electric Cipher Machine, and, on the American side, the ECM with an adapter for Type X cryptograms.
The menus that the Washington codebreakers made up were sent by pneumatic tube to the high-speed bombes on the first and second floors of a new building on the school’s 35 acres. The American bombes, built by the National Cash Register Company in Dayton, Ohio, were designed to deal with the four-rotor Enigma. Each was the equivalent of six British three-rotor bombes. They ran much faster than the early British models. These monsters were about 7 feet tall, 10 feet wide, and 2 feet deep and weighed about 2½ tons. On their gray metal faces were two rows of black disks marked 00 to 25 around their circumference; centered within each was a rotatable brass pointer. Underneath ranged four rows of eight wheels each, the rotor analogues. The machines’ vacuum tubes, relays, drive motor, clutch, brake, backup motor, and electrical losses generated so much heat that large air conditioners were needed every 10 feet, summer and winter. Once installed, the machines were operated twenty-four hours a day. They were tended by members of the navy’s women’s auxiliary—the WAVES (Women Accepted for Volunteer Emergency Service), the American counterpart of the Wrens. Trained mechanics installed and maintained the machines and made repairs—usually of broken carbon commutator brushes, of which there were hundreds on each machine.
When a WAVE received a menu, she would set up the bombe and start it running. If the crib was wrong, or if a roto
r turnover occurred during it, the bombe would complete its run in as little as 10 or 15 minutes; some runs took longer. However, when a crib was right, a hit would occur, indicating that an electrical pathway had been found. The machine ran so fast that the connection merely activated a memory circuit, which recorded the position of the hit, turned off the power, and braked the drive shaft, then reversed the rotors until they returned to the hit position. A printer typed out the rotor positions. Then the machine started again—the hit might have been caused by a chance arrangement of the rotor wiring—and ran until another hit occurred or it reached the end of the run and stopped automatically.
The hit results, called “stories,” were sent back up the pneumatic tube to the cryptanalysts, who would test each on a hand device that replicated the Enigma to see which one produced German text. To Lieutenant James T. Pendergrass, it was always a thrill to crack a cryptogram. “There’s nothing like real blood,” the cryptanalysts said to one another, referring to solving the life-or-death messages that the Germans were trying so hard to keep from them. The work, many of them felt, was exciting.
Recovered keys were turned over to the WAVES, who typed out the intercepts on Enigma analogues that turned out the plaintexts in long strings of letters. Yeomen carried these sheets of paper upstairs to the translators’ office and dropped them into a wire basket. The translators (normally, two were on duty at a time) divided the strings of letters into words and converted them into English. One of the translators, a WAVE with a doctorate in German, Erminnie Bartelmez, a short, shy, vigorous woman, felt sad when she translated a message telling some U-boat sailor that his family had been killed in an air raid, even though they were the enemy. The flow of intercepts was intermittent, and during the quiet periods the translators checked over old messages for errors or gossiped, read, or knitted. Yeomen typed up their handwritten translations and carried them to the watch officers, who brought them to the intelligence section on the second floor.
Here, in a large office with windows on one side and a balcony on the other, filled with stenographers and littered with duplicating machines, the intercepts were turned into intelligence. Like the O.I.C, OP-20-GI-2(A) maintained a file on U-boat matters. It consisted of 5-by-8-inch cards in open-topped file boxes on three or four wide shelves, where it was easy to get to. Each card carried a solved intercept reproduced by a Ditto machine in its characteristic purple ink. Each message was copied on from two to eight cards, depending on its subject matter, and the cards filed under several categories. The master file was chronological, but files also existed on individual U-boats, on U-boat position reports, on assignments, on new equipment, weather reports, status reports, and so on. Though the files primarily served the translators and the watch officers answering questions about U-boats from the American Submarine Tracking Room, they were often used by the cryptanalysts for suggestions for cribs. Once a senior watch officer, reserve Lieutenant Knight McMahan, a Ph.D. in philosophy, whose duties included overseeing the files, found just the message needed by the cryptanalysts—who came upstairs a little later all smiles.
McMahan was stunned by the events he learned about. Fresh from academe, he had had no idea of the cruelty of the sinkings or of the secret war, and he thought it was a horrible introduction to the real world. He empathized with the merchant seamen who were dying in the North Atlantic, and the images he formed of them, fed by the intercepts, motivated him in his work.
The watch officers read the day’s intercepts as they came in, saw to it that the information from them was displayed on the large maps of the Atlantic that showed U-boats and convoys, added clarifications to the messages, and talked frequently on the telephone with the tracking room, whose staff needed to know, for example, where a U-boat had been or how much fuel it had left. Among the office’s outputs was a daily report on the movement of the vessels of such neutrals as Spain and Portugal. But its main product was issued by Harvard logician and naval reserve Lieutenant Willard Van Orman Quine. He had gotten into the work after a historian of science told him that a Harvard astronomer was recruiting for cryptanalysts; Quine met with him and signed up for the navy’s correspondence course in codebreaking. Quine produced the daily U-boat summary, duplicated on paper with “Top Secret Ultra” in magenta at the head of each page.
By mid-1943, the U.S. and British cryptanalysts had developed such familiarity with the German naval signals organization that they rarely failed to find a crib, and they had enough high-speed bombes to try many cribs. From August of that year, naval Enigma was read regularly and rapidly without significant interruption for the rest of the war. This triumph, the result of hard work by brilliant people hidden in the shadows and the daring of men at sea, was the greatest extended intelligence exploit of all time.
Quine’s U-boat situation report, as well as individual solved, translated, and annotated intercepts, were sent—sometimes in double-sealed envelopes by hand of an officer, sometimes by secure teleprinter—to a restricted group of individuals in the seventh wing of the third floor of Main Navy in Washington. This long low structure, on the Mall near the Lincoln Memorial facing Constitution Avenue, had been built as a temporary office building during World War I.
On December 27, 1942, a special unit, the so-called Secret Room, had been created to handle the intercepts. The unit formed a part of the thirty-man Combat Intelligence Section of the commander in chief, United States Fleet, Admiral Ernest J. King. (In another capacity, chief of naval operations, King was in charge of the much larger Office of Naval Intelligence, which dealt with long-range and strategic issues.) The Combat Intelligence Section and its commander were designated F-2; its Atlantic branch was F-21 and its Pacific, F-22. The quiet, boyish-looking Commander Kenneth A. Knowles, who had graduated 16th out of 579 in his Annapolis class of 1927 was F-21; he had quit his naval career because of nearsightedness but had been recalled after the outbreak of war. When, on May 20, 1943, King created the Tenth Fleet as an administrative body (it had no ships) to direct the antisubmarine war, he designated Knowles’s unit as the intelligence staff for the Tenth Fleet as well as the U.S. Fleet.
F-21’s Submarine Tracking Room, itself a confidential operation, was the only area through which the Secret Room could be entered. The Secret Room, F-211, was kept locked, and only the three men who worked in it, together with Knowles and a relief officer, had keys; no one else was admitted except a few higher-ups with special permission. In charge was a New York lawyer and naval reservist, Lieutenant John E. Parsons. As the intercepts came to the Secret Room, Parsons and his assistants entered the information on a daily location list of U-boats. From here it was transferred to one of three large wall maps—the North Atlantic map was 12 feet wide—on which the U-boats were represented by pins. Each one carried the bigram assigned by the Admiralty to that submarine, the skipper’s name, the date of the position, and a colored tab to show whether it was a refueler, minelayer, cargo carrier, or combat sub. Other wall charts showed U-boats at the five Biscay bases, areas in which U-boats used particular frequencies (helpful in locating U-boats to which Dönitz sent messages), places where submarines had been sunk in the past month, and the results of OP-20-GI-2(A)’s special studies, such as one on areas that Dönitz considered especially dangerous because of aircraft. In addition, the Secret Room contained a card file of U-boats; logs of attacks on ships and on U-boats; a list of skippers; files of U-boat arrivals and departures, fuel, destroyed U-boats, and flotillas; and assessments of attacks on submarines. All of these depended on information that had been abstracted from the intercepts.
Intelligence moved out of the Secret Room in a variety of ways. Individuals, including King, viewed the charts and had the briefing officer in the room answer their questions. Every ten to fourteen days one of Parsons’s assistants prepared a summary of trends in U-boat operations as well as of new tactics and equipment; this went to the officers who had access to F-211. Messages were sent to the Admiralty about U-boat identifications, which ones wer
e involved in attacks, and similar matters. And the Secret Room staff gave the cryptanalysts information, using a special telephone for short items and notes for more complicated ones.
As head of Atlantic combat intelligence, Knowles was probably the chief consumer. His unit combined Parsons’s ULTRA intelligence with intelligence from direction-finding, prisoners of war, action reports from Allied merchantmen and warships, reconnoitering airplanes, and other sources to figure out the U-boats’ daily estimated positions. The Submarine Tracking Room had an officer and three enlisted men on watch at all times; the WAVES eventually replaced all the men except two officers not qualified physically for sea duty. (Captain Henri H. Smith-Hutton, who headed Combat Intelligence in 1943 and 1944, maintained that the plotting room ran better with the WAVES than it had earlier because the displaced men “were not as smart as these carefully selected WAVES.”) Probably Knowles himself drafted or at least approved the daily U-boat situation estimate, with its forecast of where the submarines were going, that was transmitted to task forces at sea under the signature of the commander in chief, U.S. Fleet.
Knowles’s detailed information went across the hall to the Tenth Fleet’s Convoy and Routing Section, F o. This body, the American version of the Admiralty’s Trade Division’s Movements Section, directed shipping in the U-boat-infested waters.
On a 40-foot-wide chart of the North Atlantic covering one wall of this room, U-boat positions were continually updated with pins. The section never knew that the submarine locations were furnished by codebreaking, though some suspected it; the cover story was that direction-finding provided this information. As the officers saw that “their” ships were about to encounter U-boats, they ordered diversions to detour them away from the submarine danger.
