A Tale of Two Subs Read online

  Another result of the sneak attack was a conspiratorial air of recrimination about who was responsible for the catastrophe. Although Japan’s aircraft carriers effectively ruled the entire ocean long before they even lifted anchor on a course for Pearl Harbor, the prevailing attitude in the United States among sailors and civilians alike was that sufficient warning would have somehow averted the disaster. The Navy looked to radio intelligence with renewed interest, and when Rochefort requested more personnel and more space in the Administration Building to put them, he generally got whatever resources were available. When the USS California sank at Pearl Harbor, the ship’s band no longer had a ship to play for—or any instruments to play with—so the district commandant assigned them to Hypo. Their abilities as musicians translated well into their new work as code breakers, and despite a wrongheaded attempt to wrest them from Hypo that went all the way to Admiral Kimmel’s legendary replacement, Admiral Chester Nimitz, the unlikely “crippies” (cryptographers) stayed there and made significant contributions throughout the war.

  Jasper Holmes had been delivering radio traffic summaries to the CincPac (Commander in Chief, Pacific) offices since November, and when subs began leaving for patrols in the mysterious Japanese-held Marshall and Caroline islands in December, the sub captains came to him for whatever little information he could provide for upcoming patrols. He was also briefing the submarine command about ship positions with his charts every morning. Now the collaboration with his old buddies in the sub force started to develop into what was perhaps the single most effective and devastating tactic against the Japanese navy. Since the onset of hostilities, Rochefort and Holmes had been watching the positions of Japanese submarines. The sub captains often betrayed their positions by shelling American land bases with their deck guns before the end of a patrol and were often very chatty in their radio broadcasts. In a technique called direction finding or DF, two or more Navy radio receivers at different locations would pick up the direction of these broadcasts, and determine their location by triangulation. Holmes tracked the Japanese sub I-173 across the Pacific as it shelled San Diego, then Midway. ComSubPac Tommy Withers got word that a Japanese submarine would probably be intersecting the path of the submarine USS Gudgeon at a particular point on January 18, 1942, but Holmes almost certainly didn’t tell Withers that they’d derived that from radio intelligence, or that the I-173 had been broadcasting frequently along the way.

  To his credit, Withers took the chance, and skipper Joe Grenfell of the Gudgeon described the ship when it miraculously came his way as “fat, dumb, and happy . . . the men were lounging on the upper deck, sunbathing and smoking.” Grenfell fired three torpedoes, but this upset the trim of the boat and he was unable to see what happened while his periscope dipped under the water. He heard an explosion, then silence. When he was able to get another look, the I-173 was gone, and he guessed that they may have dived with the ports open, sinking their own boat. Whether this was the case or the torpedo sank them was now academic, because their daily broadcast was never heard again, and the Japanese navy ceased references to the ship.

  The code breakers had also broken a minor code that had major implications. The Japanese navy amassed at Truk in the Caroline Islands and so Truk came to be known as Japan’s “Gibraltar of the Pacific.” The port director there broadcast announcements about the departure of ships and convoys, as well as their noontime positions as they followed a strict, prescribed path to their destinations. Rochefort may not have believed his great fortune in discovering the names, destinations, and exact courses of enemy ships from this important port. Holmes immediately recognized that it would give the submarine skippers great chances for prize targets, instead of roaming the vast Pacific looking for opportunities. In this instance, however, there was no way for Holmes to conceal how they’d come by the information. There were no coast watchers that far into the Pacific. No spy could transmit the comings and goings of ships from Truk with sufficient broadcast strength to go undetected by the Japanese. The sub commanders would get the information firsthand if it had come from a stealthy submarine; they would naturally conclude that the only way for the deskbound oddballs in Hypo to have gotten such explicit information was for them to have broken a Japanese code. When Holmes caught wind that the heavy aircraft carrier Shoho was coming out of Truk, he had no doubt that the submariners could be trusted, but Rochefort was skeptical.

  What Holmes proposed was extremely risky business. It was strictly forbidden by the guidelines for the use of radio decryption intelligence; it could not be used for minor tactical advantages, even if it was as significant as the aircraft carrier Shoho. The logic was that if the Japanese carriers suddenly had American submarines along their paths wherever they went, they would become suspicious, perhaps change the code, and leave the Americans in the dark when they most needed that source of intelligence. But as Tom Dyer said, battles are won on the basis of minor tactical advantages. Rochefort relented, but gave strict conditions: No paper was to leave the office, and Holmes would have to lie if the officers at ComSubPac questioned him closely.

  In the first of a series of surreptitious intelligence tips, Holmes took a fountain pen and wrote on his palm the map coordinates where the paths of the carrier Shoho and the submarine USS Grayling would overlap. After divulging this to his old submariner friend Charles “Gin” Styer, Withers’s chief of staff, he washed his hands thoroughly with soap and water. Styer took the information seriously and didn’t inquire too closely about its provenance. The skipper on the Grayling went to the coordinates on February 18 and followed the orders to remain submerged during the day while his soundman monitored for the Shoho. Unfortunately, he stayed too deep for periscope observations, and neglected to load the tubes in time for an attack when he got a report of heavy screws. As a result, he watched the prized target slip away.

  When ComSubPac read the skipper’s patrol report, they recognized that the opportunity for a contact with such exacting detail would never have developed without Holmes’s information, and came to think of the instance as a bird in the hand that had somehow gotten away. In his official reaction to the patrol report called “endorsements,” Tommy Withers proceeded to figuratively flay the skipper alive.

  While it’s true that the Grayling may have had a better chance of firing on the Shoho if the skipper had taken different actions, the officers at Pearl who were disappointed when decryption possibilities didn’t always translate into sinkings would discover that intercepting ships in the Pacific, as in this case, would always be a tricky business. The torpedoes had an effective range of about 2,500 yards; anything beyond that made the shot much more difficult. In order to get within 2,500 yards, the sub had to spot the target early and maneuver into position, often eluding dangerous destroyer escorts in their path. If they made contact during the day, the sub would have to do this underwater, where the subs were significantly slower. While a carrier or battleship regularly clipped along at 25 knots, a U.S. fleet boat could go only about 8 knots underwater for a half hour before its batteries were depleted. Usually it would lurk at about 2 to 3 knots to conserve its batteries for the time when it might get depth-charged—a certainty if it fired on a ship or convoy with escorts. Even the stars had to be aligned in order for any of these contacts to develop, because both navies navigated with celestial observations. If a storm was brewing up for a couple of days before a contact, the sub might not have a good celestial reference to fix its exact position. The same was true for the navigators on the target ships, and a navigation inaccuracy on the submarine compounded by an inaccuracy on the other ship could put the enemies so far away from each other at the appointed time that they might never see each other.

  Despite the many frustrations, the submarine force recognized that what Holmes gave them was gold—wherever it came from—and coordinated their plans with him whenever possible.

  During this time, the government was evacuating the dependents of military personnel from the Hawaiian Islands
, just as they had in the Philippines before the war. For Rochefort’s and Holmes’s friend John Cromwell, as well as for countless others, it meant separation from their families, sometimes forever. Like many submarine families, the Cromwells moved to the familiar territory of Palo Alto, California. It was close enough to San Francisco to be accessible to the Mare Island Navy Yard, where the Navy built and overhauled submarines, but far enough to be—at that time—less expensive than San Francisco. For John Cromwell’s son, Jack, and daughter, Ann, this meant new schools, a new neighborhood, and growing up without their father. Adding more strain to the situation was the presence of Margaret Cromwell’s parents, who had moved in with them at their home in Palo Alto. Their grandparents had fallen on hard times during the Depression and, with the onset of old age, found it difficult to find jobs and support themselves.

  While the family strained to hold itself together, Jack tried to take on the responsibility of being the man of the family—an unbidden role he tried his best to fulfill, for the time being anyway. Many of his new schoolmates found themselves in similar circumstances—an age of bubblegum and comic book heroes, of afternoon radio serials and matinees at the Bijou. Newsreels like The March of Time struck people differently in that time of intense patriotism; it seemed that those who had no family in the war thought of individuals’ sacrifices with a detached, almost uncomprehending mien. Nevertheless, the propaganda spurred people on to join the armed services, buy war bonds, make Victory Gardens, and recycle war materials like steel cans and tires. They also swallowed tax increases as well as the rationing of nearly everything: gasoline, meat, butter, even bread. It didn’t matter how much money you had—if you didn’t have the coupons to buy certain commodities, you just couldn’t buy them. The newsreels were persuasive, shocking, effective, and for those who did have family members on the front lines, they were frightening. Young Ann Cromwell frequently cried.

  Back in Hawaii, the military couldn’t enforce the evacuation among personnel with families classified as residents, who owned homes and paid mortgages, like Holmes and Dyer. The families who decided to stay found their social lives suddenly truncated as many of their civilian friends moved back stateside and the Navy built Pearl Harbor back, bigger even than anyone had ever imagined. Honolulu and the surrounding area also went through major changes as the Hawaiian Islands became the key staging area for fleet movements and invasions. Despite having the comforts of home, due to the nature of his decryption work Dyer would leave home with a lunchbox full of sandwiches his wife packed for him, and wouldn’t return for days at a stretch, until he ran out of food or Rochefort ordered him to take a break and get lost for a couple of days.

  All of them worked around the clock in the basement, sometimes sleeping fitfully on cots in the hallways for a few hours like college students cramming for a test. But the consequences of their test were literally a matter of life and death, and as the cryptographers grappled with the befuddling nuances of the JN-25 code, the cool, dank atmosphere in the basement of the Administration Building took its toll in the form of colds that often developed into pneumonia. Rochefort seldom left the building and took to wearing a red smoking jacket and carpet slippers, which gave rise to the legend that he was a bizarre eccentric. While this attire may not have been strictly Navy regulation and it was true that he had a strong personality that sometimes got him in trouble, he was by no means an eccentric. The mad professor aura created by the smoking jacket and carpet slippers was really just a pragmatic attempt to keep warm, but somehow it stuck and he is remembered that way in books and films. He was, on the other hand, a very dedicated commander who marshaled his forces to the point of exhaustion but no further. He gave them considerable latitude and loyally backed them up in bureaucratic dustups while he and his crew were accomplishing a half year’s work in a matter of weeks.

  The initial break on JN-25 came from Lieutenant Rudolph J. Fabian, a cryptologist at the Navy’s radio intelligence unit at Corregidor (code name Cast) in the beseiged Philippines. The main OP-20-G headquarters in Washington, D.C., also called “Negat,” worked on the codes as well, but it was Rochefort’s group that was able to make most of the breaks.

  Coding is the substitution of a letter, word, number, or concept with an unrelated signifier. For instance, we could code the letters a, b, c, t, and u as a=APPLE, b=BALL, c=CHARLIE, t=TOP, u=UNICORN. Working backward, the decode for BALL-APPLE-TOP would be the word “bat.” If the coding is randomized, the code becomes more secure. For example, the sender and receiver of a code may agree to shift the letters in the code up a value of one on a predetermined day. In that case, a=BALL, b=CHARLIE, t=UNICORN, and the decode for CHARLIE-BALL-UNICORN would be the word “bat.” This method is somewhat less intuitive than the original code, and would be slightly more difficult for an amateur. But it is susceptible to a trained cryptologist because of the similarity in values between “a” (apple/ball) and “b” (ball/Charlie). A skilled cryptologist will immediately recognize the pattern using innate skill, mathematical analysis, and intuition to “attack” such a code and break its secrets.

  The way to get around this security problem is to use encryption, which is distinct from encoding. The simplest encryption is to substitute letters or words for numbers (encoding), then manipulate the numbers in some way to obscure their meaning (encryption). For instance, we might encode the alphabet thus: a=1, b=2, c=3, and so on. Then we might encrypt it by squaring the numbers in the code: a=12=1, b=22=4, c=32=9, and so on. Therefore, the decrypt of the cipher text 9-1-4 would be the square roots of the sequence, 3-1-2, and the decode of 3-1-2 would be the word “cab.”

  The encryption, or scrambling, of the numbers could be quite elaborate. However, given enough examples of a certain code, an astute cryptographer can divine even elaborate computations because they betray themselves by repeating a certain pattern. Overly complicated computations were also undesirable for two other reasons: In the days before computers, doing these calculations by hand would be more prone to error, and their unwieldiness made them too time-consuming. The Germans got around this by using a combination of electrical and mechanical randomness settings on a machine called Enigma. Enigma could be described as an analog computer to automate the encryption and decryption of signals, and used a series of rotors that rotated around a common axis. Essentially, it was a series of magic decoder rings, and each successive ring made the encryption exponentially more difficult to decrypt. A cryptographer trying to break the code would need not only an Enigma machine, but also the initial rotor settings and plugboard settings in order to read the same text as the recipient.

  All of this increased the complexity of the encryption and the apparent randomness, and it decreased the amount of work to encrypt and decrypt. Enigma was, however, a machine that repeated certain patterns. Every rotor was wired to scramble in a certain, discrete way, and the Enigma machine’s Achilles’ heel was that the subsequent rotors relied on the motion of the first one, in what was a very predictable pattern. With initial help from expatriate Polish code breakers about the design of the Enigma machine, the British code breakers at Bletchley Park were able to discern those patterns. The British created a method to reduce the number of possible permutations, thereby eventually cracking the code. The Americans mechanized this attack method to determine the initial rotor and plugboard settings, and began decrypting Enigma transmissions in earnest.

  In popular culture, the Germans get much credit for using such a diabolically clever device, and the British get even more for cracking its codes. It is interesting to note that the brilliant cryptographer William Friedman in the U.S. Army’s Signals Intelligence Service (SIS), had a similar, though superior, concept at about the same time. Friedman added more rotors and also recognized the vulnerability of the rotor stepping mechanism, so he added a feature that stepped the rotors in a somewhat more random pattern.* The machine he created was called the SIGABA, and the Navy’s version was called the ECM-2. At the beginning of the war, submarin
es like the Sculpin had to observe the “hundred fathom curve,” a line they could not cross in enemy territory because if the sub sank in anything shallower than a hundred fathoms, or 600 feet, it could conceivably be salvaged, thereby compromising the ECM-2’s secrecy. Fortunately, its code was never broken during the war, and was only retired in 1959, when faster alternatives became available.

  In the 1930s, Friedman’s team also cracked the Japanese diplomatic code they called Purple. Years before, the cryptologists went the simplest route when a rapid solution was desired: They cheated by stealing a copy. The SIS had the FBI perpetrate a black bag job—essentially cat burglary—at a Japanese diplomatic office to get the Japanese diplomatic codes. The FBI photographed the codebooks and replaced them without the diplomatic staff suspecting anything. A former missionary to Japan undertook the painstaking translation of the code, and while he did he kept it in a red binder, so the code became known as the Red code. The precedent having been set, subsequent codes were also referred to with color designations: Jade, Coral, and eventually Purple.

  Like Germany’s Enigma and America’s ECM-2, Purple was transmitted by an encryption/decryption device that the Japanese called the Type 97. It used stepped rotors, but rather than using rotors with twenty-six steps for the twenty-six characters in the alphabet, it used two ten-step rotors for the consonants and a six-step rotor for the vowels A, E, I, O, U, and Y. This added a level of complexity that proved difficult to solve. Nevertheless, the underlying patterns betrayed the machine’s vulnerability, and the code breakers at SIS were able to ascertain the patterns and break the code. Going one step further, and with the help of SIS cryptographer Leo Rosen, Friedman built a Purple machine so that they wouldn’t have to decrypt each message by hand. To start with an unintelligible string of characters and end with a machine to turn them into intelligence gold was a remarkable achievement. Sadly, the strain and long hours of breaking Purple and making a machine took its toll on Friedman, and after having given so much to protect his country, he had a nervous collapse. He recuperated somewhat and would continue to contribute, but only in a reduced capacity for limited hours.